Oxycodone for cancer‐related pain

Summary of findings 1. CR oxycodone compared with IR oxycodone for cancer‐related pain in adults

Outcomes	*Illustrative comparative risks (95% CI)**		Relative effect (95% CI)	No of participants (studies)	Certainty of the evidence (GRADE)	Comments
CR oxycodone compared with IR oxycodone for cancer‐related pain in adults
Patient or population: adults with cancer‐related pain Settings: in‐ or outpatients Intervention: CR oxycodone Comparison: IR oxycodone
	Assumed risk	Corresponding risk
	IR oxycodone	CR oxycodone
Pain intensity (mean across treatment or at end of treatment; length of treatment varied across trials; various pain intensity scales; SMD)	The mean pain intensity in the CR oxycodone group was 0.12 standard deviations higher (0.1 lower to 0.34 higher) than in the IR oxycodone group		SMD 0.12 (‐0.1 to 0.24)	319 (3 studies)	⊕⊕⊝⊝ very low^a,b	‐
Constipation (Event rate during treatment, length of treatment varied across trials)	224 per 1000	159 per 1000 (101 to 253)	RR 0.71 (0.45 to 1.13)	317 (3 studies)	⊕⊝⊝⊝ very low^c,d	‐
Drowsiness/somnolence (Event rate during treatment, length of treatment varied across trials)	224 per 1000	230 per 1000 (154 to 344)	RR 1.03 (0.69 to 1.54)	317 (3 studies)	⊕⊝⊝⊝ very low^c,d	‐
Nausea (Event rate during treatment, length of treatment varied across trials)	242 per 1000	206 per 1000 (136 to 310)	RR 0.85 (0.56 to 1.28)	317 (3 studies)	⊕⊝⊝⊝ very low^c,d	‐
Vomiting (Event rate during treatment, length of treatment varied across trials)	174 per 1000	115 per 1000 (66 to 200)	RR 0.66 (0.38 to 1.15)	317 (3 studies)	⊕⊝⊝⊝ very low^c,d	‐
Quality of life	No data available, but there appeared to be no difference in treatment acceptability between the treatment groups (measured on various scales, not pooled); 578 participants (3 studies); quality of the evidence low.^a
Participant preference	No data available		‐	‐	‐	‐
The assumed risk* is reported as the observed risk in the control group across studies. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; CR: controlled‐release; IR: immediate‐release; RR: risk ratio; SMD: standardised mean difference.
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate; the true effect is likely to be close to the estimate of effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited; the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate; the true effect is likely to be substantially different from the estimate of effect.
^a Downgraded twice for very serious limitations to study quality due to risk of bias (attrition bias and under‐reporting of the domain of selection bias) ^b Downgraded once for imprecision due to low event rates ^c Downgraded twice for imprecision due to very low event rates ^d Downgraded twice for very serious limitations to study quality due to risk of bias (performance/detection bias, and inadequate titration and under‐reporting of the domains of selection, performance, detection and attrition bias, and whether the participants were adequately titrated)

See: Summary of findings 1 CR oxycodone compared with IR oxycodone for cancer‐related pain in adults; Summary of findings 2 CR oxycodone compared with CR morphine for cancer‐related pain in adults

Summary of findings 2. CR oxycodone compared with CR morphine for cancer‐related pain in adults

Outcomes	*Illustrative comparative risks (95% CI)**		Relative effect (95% CI)	No of participants (studies)	Certainty of the evidence (GRADE)	Comments
CR oxycodone compared with CR morphine for cancer‐related pain in adults
Patient or population: adults with cancer‐related pain Settings: in‐ or outpatients Intervention: CR oxycodone Comparison: CR morphine
	Assumed risk	Corresponding risk
	CR morphine	CR oxycodone
Pain intensity (various pain intensity scales; SMD)	The mean pain intensity in the CR oxycodone group was 0.14 standard deviations higher (0.01 lower to 0.27 higher) than in the CR morphine group		SMD 0.14 (0.01 to 0.27)	882 (7 studies)	⊕⊕⊝⊝ low^a	This estimate did not include the data from the Chinese language studies (n = 2) as sensitivity analyses indicated they were not comparable. Converting the SMD as a difference in Brief Pain Inventory scores (0 to 10 numerical rating scale from no pain to worst pain imaginable) between the treatments gave an estimated difference of 0.27 between the treatments, which was not clinically significant.
Constipation ‐ all available data (Event rate during treatment, length of treatment varied across trials)	322 per 1000	241 per 1000 (212 to 277)	RR 0.75 (0.66 to 0.86)	1894 (18 studies)	⊕⊕⊝⊝ low^b	This estimate did include the data from the Chinese language studies.
Constipation ‐ only English language studies (Event rate during treatment, length of treatment varied across trials)	355 per 1000	348 per 1000 (291 to 412)	RR 0.98 (0.82 to 1.16)	797 (5 studies)	⊕⊝⊝⊝ very low^b,c	This estimate did not include the data from the Chinese language studies (n = 13) as sensitivity analyses indicated they were not comparable.
Drowsiness/somnolence (Event rate during treatment, length of treatment varied across trials)	228 per 1000	201 per 1000 (169 to 239)	RR 0.88 (0.74 to 1.05)	1486 (15 studies)	⊕⊕⊝⊝ low^b	This estimate did include the data from the Chinese language studies as sensitivity analyses indicated they were comparable.
Nausea (Event rate during treatment, length of treatment varied across trials)	231 per 1000	215 per 1000 (178 to 259)	RR 0.93 (0.77 to 1.12)	1388 (13 studies)	⊕⊕⊝⊝ low^b	This estimate did include the data from the Chinese language studies as sensitivity analyses indicated they were comparable.
Vomiting (Event rate during treatment, length of treatment varied across trials)	157 per 1000	127 per 1000 (99 to 163)	RR 0.81 (0.63 to 1.04)	1388 (13 studies)	⊕⊝⊝⊝ very low^b,c	This estimate did include the data from the Chinese language studies as sensitivity analyses indicated they were comparable
Quality of life	No data available, but CR oxycodone appeared to be associated with similar or lower treatment acceptability than CR morphine (measured on various scales, not pooled); 149 participants (3 studies); quality of the evidence verylow^d,e					‐
Participant preference (end of treatment in a cross‐over trial with each phase lasting seven days)	8/23 participants preferred CR oxycodone while 11/23 participants preferred treatment with CR morphine.		‐	23 (1 study)	⊕⊝⊝⊝ very low^d,e	‐
The assumed risk* is reported as the observed risk in the control group across studies. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; CR: controlled‐release; RR: risk ratio; SMD: standardised mean difference.
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate; the true effect is likely to be close to the estimate of effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited; the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate; the true effect is likely to be substantially different from the estimate of effect.
^a Downgraded twice for very serious limitations to study quality due to risk of bias (performance/detection bias and attrition bias and under‐reporting of the domain of selection bias) ^b Downgraded twice for very serious limitations to study quality due to risk of bias (performance/detection bias and/or attrition bias and under‐reporting) ^c Downgraded once for imprecision due to low event rates/participant numbers ^d Downgraded twice for imprecision due to very low event rates/participant numbers ^e Downgraded twice for very serious limitations to study quality due to risk of bias (attrition bias and under‐reporting of the domain of selection bias)

Background

This review is an update of a previously published review in the Cochrane Database of Systematic Reviews, 2017, Issue 8 on oxycodone for cancer‐related pain.

Description of the condition

Pain from cancer can be caused by direct invasion of a tumour into soft tissue or bone and is often a presenting symptom at the time of diagnosis of cancer. One European survey published in 2009 found that, of 5000 people with cancer (including 617 community‐based National Health Service (NHS) patients in the UK), 72% experienced pain (77% of UK patients) which was of moderate to severe intensity in 90% of this group (Breivik 2009). This is consistent with a systematic review that demonstrated cancer pain prevalence of up to 75% in advanced disease, and that at least 30% of people with cancer are undertreated (Greco 2014). Recent research has also shown that less than half of all people with cancer that die are prescribed a strong opioid, and that median treatment duration is only nine weeks before death (Ziegler 2016). Pain in people with cancer may also be caused by cancer treatments and by comorbid conditions. In this review, we define cancer pain as pain arising as a direct consequence of the cancer, and not from other aetiologies.

Description of the intervention

Oxycodone is a strong opioid analgesic indicated for the treatment of moderate to severe chronic pain, including cancer pain. It is available orally as immediate‐release (IR) solution and tablets (for four‐hourly dosing) and as sustained (controlled (CR)) release tablets (for 12‐hourly dosing). It is also available as a parenteral injection. In some countries, oxycodone is available as a compound with paracetamol (acetaminophen) or ibuprofen.

How the intervention might work

Oxycodone works primarily as an agonist of mu‐opioid receptors in the spinal cord and brain. It has some activity at kappa‐opioid receptors (which are also involved in nociception or analgesia) though the importance of this mechanism in the overall analgesic effect of oxycodone is unclear. Despite animal studies suggesting differences in pharmacodynamics, these have not been demonstrated in clinical studies to date. Therefore, the shared mechanism of action to other strong opioids (i.e. agonist activity at mu‐opioid receptors) means that clinical benefits and adverse effects are likely to be similar. However, important differences exist in the pharmacokinetics of strong opioids (e.g. morphine undergoes second‐phase elimination via glucuronidation, while oxycodone undergoes extensive first‐phase metabolism via CYP2D6 and CYP3A4 pathways) so clinical equivalence cannot be inferred (Gudin 2012; Leppert 2010).

Why it is important to do this review

The World Health Organization (WHO) published the Method for Cancer Pain Relief (WHO analgesic ladder) in 1986 (WHO 1986), which advocates a stepwise approach to analgesia for cancer pain and revolutionised the use of oral opioids. It recommended that morphine be used first‐line for moderate to severe cancer pain. Over 30 years on, WHO guidelines continue to support opioids for moderate to severe cancer pain management (WHO 2018). Observational studies have suggested that this approach results in pain control for 73% of people (Bennett 2008) with a mean reduction in pain intensity of 65% (Ventafridda 1987).

Many people with cancer experience moderate to severe pain that requires treatment with strong analgesics. Oxycodone and morphine are examples of strong opioids that are used for the relief of cancer pain. However, strong opioids are not effective for pain in all people, neither are they well tolerated by all people. Guidance by the European Association for Palliative Care on the use of opioids in cancer pain suggests that oxycodone could be used as first‐line treatment of moderate to severe cancer pain as an alternative to morphine (Caraceni 2012). The aim of this review is to assess whether oxycodone is associated with better pain relief and tolerability than other analgesic options for people with cancer pain. The protocol for this review was updated from Reid 2010.

Objectives

To assess the effectiveness and tolerability of oxycodone by any route of administration for pain in adults with cancer.

Methods

Criteria for considering studies for this review

Types of studies

Randomised trials are the best design to minimise bias when evaluating the effectiveness of an intervention. We included randomised controlled trials (RCTs), with parallel‐group or cross‐over design, comparing oxycodone (any formulation and any route of administration) with placebo or an active drug (including oxycodone) for cancer background pain. We did not examine studies on breakthrough pain.

Types of participants

Adults (aged ≥ 18 years) with cancer pain. We did not restrict inclusion by cancer type or body region affected, but rather included adults with pain from any type of cancer in any body region.

Types of interventions

Oxycodone (any dose/frequency, formulation, and route of administration) versus oxycodone (any other dose/frequency, formulation, and route of administration)

Oxycodone (any dose/frequency, formulation, and route of administration) versus other active drug (any dose/frequency, formulation, and route of administration)

Oxycodone (any dose/frequency, formulation, and route of administration) versus placebo

Types of outcome measures

Primary outcomes

Pain was the primary outcome. This is often reported as pain intensity (typically reported as a continuous measure, measured on a visual assessment scale (VAS) or numerical rating scale (NRS)) or pain relief (typically reported as a categorical measure), which we have treated as measures of essentially the same thing, albeit from opposite starting points (i.e. effectiveness is demonstrated by pain intensity going down and by pain relief going up). The majority of the included studies reported pain intensity, with much fewer studies reporting pain relief apart from the newly added Chinese‐language studies (see also Subgroup analysis and investigation of heterogeneity; Sensitivity analysis; Excluded studies). We therefore meta‐analysed pain intensity and treated this as our primary outcome measure (see also Measures of treatment effect; Data synthesis), but we also meta‐analysed pain relief, where possible.

Both these outcomes had to be participant‐reported and could be reported in any transparent manner (e.g. by using numerical or verbal rating scales). We did not consider these outcomes when reported by physicians, nurses, or carers. If possible, we aimed to distinguish between nociceptive and neuropathic pain, but the data were not presented in a manner that made this possible.

Secondary outcomes

Adverse events (e.g. constipation, nausea, vomiting, drowsiness, confusion, respiratory depression).
Quality of life (or treatment acceptability as a proxy).
Participant preference.

We considered all these outcomes as they were reported in the included studies.

Search methods for identification of studies

We applied no language, date, or publication status (published in full, published as abstract, unpublished) restrictions to the search.

Electronic searches

For this update, we identified relevant trials by searching the following databases:

Cochrane Central Register of Controlled Trials (CENTRAL) in the Cochrane Library (Issue 11 of 12, 2021);
MEDLINE and MEDLINE In‐Process (Ovid) (Nov 2016 to 24 November 2021);
Embase (Ovid) (Nov 2016 to 2021 4 November);
Science Citation Index (Web of Science) (2016 to 29 November 2021);
Conference Proceedings Citation Index ‐ Science (Web of Science) (2016 to 29 November 2021);
BIOSIS (Web of Science) (2016 to 30 November 2021);
PsycINFO (EBSCO) (2016 to November 2021).

We applied the Cochrane highly sensitive search strategy for identifying RCTs to this search (Lefebvre 2021). When these searches were run for the original review in March 2014, PubMed was also searched. We did not search PubMed for the subsequent updates as it did not yield any records that were not found by the other databases in the original review. The search strategies used can be found in Appendix 1.

Searching other resources

For this update, we ran searches on 30 November 2021, on Clinicaltrials.gov, EU Clinical Trials Register, WHO International Clinical Trials Registry Platform (ICTRP), and UK Clinical Trials Gateway (UKCTG), but not metaRegister of Controlled Trials (mRCT) as this is no longer available, but rather signposted to EU Clinical Trials Register and UK Clinical Trials Gateway (UKCTG). For both the previous review and the updates, we checked the bibliographic references of relevant identified studies to find additional trials not identified by the electronic searches and contacted authors of the included studies to ask if they knew of any other relevant studies.

Data collection and analysis

Selection of studies

Two of three review authors (MSH, NB, AJP) assessed the titles and abstracts of all the studies identified by the search for potential inclusion. We independently considered the full records of all potentially relevant studies for inclusion by applying the selection criteria outlined in the Criteria for considering studies for this review section. We resolved any disagreements by discussion. We did not restrict the inclusion criteria by date, language, or publication status (published in full, published as abstract, unpublished).

Data extraction and management

Using a standardised, piloted data extraction form, two review authors (MSH, JSH) extracted data pertaining to study design, participant details (including age, cancer characteristics, previous analgesic medication, and setting), interventions (including details about titration), and outcomes. We resolved any disagreements by discussion. If there were studies for which only a subgroup of the participants met the inclusion criteria for the current review, we would have only extracted data on this subgroup provided randomisation had not been broken; however, no such studies were identified for inclusion.

Assessment of risk of bias in included studies

Two review authors (MSH and JSH or NB or YC) independently assessed the methodological quality of each of the included studies by using the risk of bias assessment method outlined in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). For each study, we assessed the risk of bias for the following domains: selection bias (study level; two items; random sequence generation and allocation concealment), performance bias (outcome level; two items; blinding of participants and blinding of treating personnel), detection bias (outcome level; one item; blinding of outcome assessment), attrition bias (outcome level; one item; incomplete outcome data), and reporting bias (study level; one item; selective reporting). For the question on allocation concealment, in the absence of explicit explanation about allocation concealment, we accepted 'central randomisation' as sufficient indication of adequate allocation concealment and consequently rated such studies as having low risk of bias for this item. We also included an item that assessed the adequacy of titration (with judgements made based on any available relevant information, including design features, inclusion criteria, and interim pain assessments) and another item that captured whether data were available for both time periods in cross‐over trials; we also listed under 'Other bias' any other biases identified, including carry‐over effects for cross‐over trials. Each of the risk of bias items required a 'low risk,' 'high risk,' or 'unclear risk' response. We also documented the reasons for each response in accordance with Higgins 2011, and resolved any disagreements on the risk of bias ratings through discussion. For the item assessing whether data were available for both time periods in cross‐over trials, we inputted 'unclear' and 'not applicable' as the rating and reason for parallel‐group trials. Finally, we also extracted and reported whether a study was free from commercial funding.

Measures of treatment effect

For pain intensity, we extracted the means and standard deviations (SDs) and we used these to estimate the standardised mean difference (SMD) between the treatments along with the 95% confidence interval (CI), as the outcome was not measured on the same scale across studies. For this update, the majority of the 18 new studies did not report pain intensity as a continuous measure, but rather pain relief as a categorical measure. In order to be able to meta‐analyse the pain data from these studies, we therefore extracted the event rates in each of these pain relief categories also to calculate and analyse them as risk ratios (RRs) with 95% CIs. For adverse events, we extracted event rates to calculate RR with 95% CIs as the summary estimates (see also Data synthesis).

Unit of analysis issues

The participant was the unit of analysis but, in a number of cases, the pain intensity data reported as a continuous measure in the included cross‐over trials could not otherwise be incorporated into the analyses (see Dealing with missing data), so we included them as if the design had been parallel‐group. Higgins 2011 (in Chapter 16) pointed out that this approach, while giving rise to unit of analysis errors, is nevertheless conservative as it results in an underweighting of the data. However, in order to assess the impact of this strategy, we also performed sensitivity analyses when we included cross‐over trial data in this manner by excluding the cross‐over trials from the meta‐analyses. We did not include dichotomous pain relief or adverse event data from cross‐over trials in this manner, but rather reported them per study. Where a study with more than two intervention arms was included, only data from intervention and control groups that met the eligibility criteria were extracted. We included studies with more than two eligible intervention arms in separate comparisons to avoid the double‐counting of participants.

Dealing with missing data

In cases where data were missing, we contacted the authors to request the missing data. This strategy did not result in any additional data. We limited imputation of missing data to the imputation of missing SDs, either by calculating the SD if enough information was available or by using SDs from similar samples or studies, both according to the methods outlined by Higgins 2011. We only imputed SDs for pain intensity for Lux 2014, which were not reported for the subgroup of participants with malignant pain, by using the reported SDs for the whole sample of participants with either malignant (n = 31) or non‐malignant pain (n = 15), and for Yu 2014, for the primary outcome of the study "mean pain at its worst in the past 24 hours", by using the standard deviations for the same outcome measured at baseline in the full analysis set. We recorded the dropout/missing data rates in the risk of bias tables under the items on attrition bias and in the 'Participants' section of the Characteristics of included studies table, and we addressed the potential effect of the missing data on the results in the Discussion. It was not possible to assess the impact of missing data in sensitivity analyses due to the low number of included studies within each comparison that were not subject to attrition bias. In all cases, we aimed to perform intention‐to‐treat (ITT) analyses.

Assessment of heterogeneity

We quantified heterogeneity by using the I² statistic. We considered I² values above 50% to represent substantial heterogeneity in line with Higgins 2011, and we planned to assess potential sources of heterogeneity through subgroup analyses as outlined in Subgroup analysis and investigation of heterogeneity. See also Data synthesis.

Assessment of reporting biases

In addition to implementing the comprehensive search strategy outlined in the Search methods for identification of studies section, the risk of outcome reporting bias was illustrated in the risk of bias summary figures that we constructed for each study and each type of assessed bias.

Data synthesis

We entered the data extracted from the included studies into Review Manager 5 (RevMan 2014), which was used for data synthesis. We analysed pain intensity using the generic inverse variance method in accordance with Higgins 2011. However, given the limitations of this analysis strategy as outlined in the Unit of analysis issues section, we also considered the results of the individual studies. We meta‐analysed the pain relief and adverse events data by using the Mantel‐Haenszel method; however, as this method is not suitable for cross‐over trial data, we only included the data from parallel‐group trials in these analyses. In addition, we have also presented all reported adverse events from the included studies in tables. As we have assumed that there is a single common intervention effect which we are aiming to estimate, we used a fixed‐effect model in all analyses. However, we did not pool the data if the I² statistic was above 50%, although we note that in some instances the pooled estimate of analyses where the I² statistic exceeds 50% will appear on the presented figures or analyses due to the way we have presented the results (in comparison‐based subgroups for each outcome) as Review Manager does not allow for the selective presentation of pooled estimates within only a subset of subgroups. In such cases, we have clearly stated in the text that the pooled estimates should be disregarded. We reported the pain intensity and pain relief data for the included studies that could not be meta‐analysed narratively, along with any other outcome data that could not be meta‐analysed (such as quality of life data).

Subgroup analysis and investigation of heterogeneity

Different aspects of the trials are likely to contribute heterogeneity to the proposed main analyses. If there were sufficient data, we therefore planned to perform subgroup analyses based on doses, titration, formulations (e.g. IR, sustained‐release), routes of administration (e.g. oral, rectal), length of the trials, and populations (e.g. opioid‐naive participants). We grouped the studies by formulation and route of administration, but as there were insufficient data, we were unable to perform any further subgroup analyses. As outlined and discussed in detail in the Excluded studies section, for this update, we performed additional sensitivity analyses that assessed the impact of including Chinese language studies.

Sensitivity analysis

If sufficient data had been available, we planned to examine the robustness of the meta‐analyses by conducting sensitivity analyses using different components of the risk of bias assessment, particularly those relating to whether allocation concealment and blinding were adequate. We also planned to conduct further sensitivity analyses to examine the impact of missing data on the results if a large proportion of the studies were at an 'unknown' or 'high risk' of attrition bias and, finally, we planned to use sensitivity analyses to examine whether publication status and trial size influenced the results. Unfortunately, we were unable to perform any such sensitivity analyses due to the low number of studies within each comparison. As already outlined in Unit of analysis issues, we performed sensitivity analyses when we included cross‐over trial data in analyses as if their designs were parallel‐group by excluding the cross‐over trials from the meta‐analyses and, as outlined in Excluded studies, we also performed additional sensitivity analyses to assess the impact of the inclusion of Chinese language studies in the analyses of the CR oxycodone versus CR morphine comparison.

Summary of findings and assessment of the certainty of the evidence

We used the GRADE system to rank the certainty of the evidence using the GRADEprofiler Guideline Development Tool software (GRADEpro GDT 2015), and the guidelines provided in Chapter 12.2 of the CochraneHandbook for Systematic Reviews of Interventions (Higgins 2011).

The GRADE approach uses five considerations (study limitations, consistency of effect, imprecision, indirectness, and publication bias) to assess the certainty of the body of evidence for each outcome. The GRADE system uses the following criteria for assigning grade of evidence:

high: we are very confident that the true effect lies close to that of the estimate of the effect;
moderate: we are moderately confident in the effect estimate; the true effect is likely to be close to the estimate of effect, but there is a possibility that it is substantially different;
low: our confidence in the effect estimate is limited; the true effect may be substantially different from the estimate of the effect;
very low: we have very little confidence in the effect estimate; the true effect is likely to be substantially different from the estimate of effect.

We decreased the grade rating by one (‐1) or two (‐2) if we identified:

serious (‐1) or very serious (‐2) limitations to study quality based on a qualitative assessment of the extent the body of evidence was at risk of bias;
important inconsistency (‐1) if the I² was above 50%. We did not downgrade twice for this domain;
some (‐1) or major (‐2) uncertainty about directness based on a qualitative assessment of the extent the body of evidence met the inclusion criteria;
imprecise (‐1) or very imprecise data (‐2) based on number of participants/events. For dichotomous outcomes, we downgraded once or twice if the total number of events was below 300 or 150, respectively, and for continuous outcomes we downgraded once or twice if the total number of participants was below 400 or 200, respectively;
high probability of reporting bias (‐1) based on a qualitative assessment of the likelihood of reporting bias. We did not downgrade twice for this domain.

'Summary of findings' tables

We included two summary of findings tables to present the main findings in a transparent and simple tabular format. In particular, we included key information concerning the certainty of evidence, the magnitude of effect of the interventions examined, and the sum of available data on the outcomes pain intensity, adverse events (constipation, drowsiness/somnolence, nausea, and vomiting), quality of life (or treatment acceptability as a proxy) and participant preference.

See Schmidt‐Hansen 2013 for the published protocol for this review.

Results

Description of studies

Results of the search

The updated searches of the electronic databases retrieved 817 records (see Electronic searches). Our searches of the trials registers identified 46 further studies. Our screening of the reference lists of the included publications revealed one additional potentially relevant study, and the search identified a systematic review of Chinese studies (Zhou 2020) with 16 potentially relevant studies. We therefore had a total of 880 records.

Once duplicates had been removed, we had 616 records. We excluded 564 records based on titles and abstracts. We obtained the full text of the remaining 52 records. We excluded 14 studies for this update. We identified 10 new ongoing studies for this update and there were three studies we could not classify so they are listed under 'studies awaiting classification'. Two of the records were additional references for Corli 2016.

From the search, we included a total of 19 new studies reported in 23 references for this update (Cao 2015; Gao 2012; Inoue 2017; Inoue 2018; Lee 2017; Li 2013; Liu 2021; Nosek 2017; Ren 2012; Song 2015; Sun 2013; Tu 2015; Wang 2008; Xie 2018; Ye 2012; Yu 2007; Yu 2009; Zhang 2011; Zhang 2016a), of which one had originally been included under 'ongoing studies'. For a further description of our screening process, see the study flow diagram (Figure 1).

Figure 1

Study flow diagram

In total, we included 42 studies (see Characteristics of included studies) and excluded 77 studies (see Characteristics of excluded studies) in this review. There are nine studies awaiting classification (see Characteristics of studies awaiting classification), and 18 ongoing studies (see Characteristics of ongoing studies).

Included studies

The 19 new studies included an additional 1836 randomised participants, such that the 42 included studies enrolled/randomised 4485 participants (2347 men, 1869 women; for the remaining 268 participants, gender was not specified) with 3945 of these analysed for efficacy and 4176 for safety. The reported mean/median ages of the participant populations in the studies ranged from 45 years to 75.3 years. Eleven of the studies were cross‐over trials (Beaver 1978a; Beaver 1978b; Bruera 1998; Gabrail 2004; Hagen 1997; Heiskanen 1997; Kalso 1990; Lauretti 2003; Leow 1995; Lux 2014; Stambaugh 2001), and the remainder were parallel‐group trials, with eight of the studies conducted in the USA (Beaver 1978a; Beaver 1978b; Gabrail 2004; Kaplan 1998; Mucci‐LoRusso 1998; Parris 1998; Salzman 1999; Stambaugh 2001); two in Canada (Bruera 1998; Hagen 1997); two in Finland (Heiskanen 1997; Kalso 1990); 18 in China (Cao 2015; Gao 2012; Li 2013; Liu 2021; Ren 2012; Song 2015; Su 2015; Sun 2013; Tu 2015; Wang 2008; Xie 2018; Ye 2012; Yu 2007; Yu 2009; Yu 2014; Zhang 2011; Zhang 2014; Zhang 2016a); three in Italy (Corli 2016; Mercadante 2010; Zecca 2016), two in Japan (Inoue 2017; Inoue 2018); and one each in Germany/Poland/Switzerland (Lux 2014), Australia (Leow 1995), Brazil (Lauretti 2003), Poland (Nosek 2017), Korea (Lee 2017), the UK (Riley 2015), and Japan/Korea (Imanaka 2013). The length of the trials ranged from single‐dose treatment to one year, and the studies reported the following comparisons:

CR oxycodone versus IR oxycodone (Kaplan 1998; Parris 1998; Salzman 1999; Stambaugh 2001);
CR oxycodone versus extended‐release (ER) oxycodone (Lux 2014);
CR oxycodone versus CR morphine (Bruera 1998; Cao 2015; Corli 2016; Gao 2012; Heiskanen 1997; Lauretti 2003; Li 2013; Mercadante 2010; Mucci‐LoRusso 1998; Nosek 2017; Ren 2012; Riley 2015; Song 2015; Sun 2013; Tu 2015; Wang 2008; Xie 2018; Ye 2012; Yu 2007; Yu 2009; Zecca 2016; Zhang 2011; Zhang 2014; Zhang 2016a), with two of these studies including a further two arms of transdermal (TD) buprenorphine and TD fentanyl (Corli 2016; Nosek 2017); and two of the studies comparing two different brands of slow‐release morphine to CR oxycodone (Zhang 2014; Zhang 2016a);
CR oxycodone versus CR hydromorphone (Hagen 1997);
CR oxycodone versus ER hydromorphone (Inoue 2017; Yu 2014);
CR oxycodone versus ER oxymorphone (Gabrail 2004);
CR oxycodone versus ER tapentadol (Imanaka 2013);
CR oxycodone versus TD fentanyl (Corli 2016; Nosek 2017; Su 2015);
CR oxycodone versus TD buprenorphine (Corli 2016; Nosek 2017);
CR oxycodone versus oral ibuprofen (Liu 2021)
IR oxycodone versus IR hydromorphone (Inoue 2018);
intravenous (IV) oxycodone versus rectal oxycodone (Leow 1995);
IV oxycodone versus IV morphine (Lee 2017);
IV oxycodone followed by IR oxycodone versus IV morphine followed by IR morphine (Kalso 1990);
intramuscular (IM) oxycodone versus oral oxycodone (Beaver 1978a);
IM oxycodone versus IM morphine versus IM codeine (Beaver 1978b).

See also Characteristics of included studies table for further details about the studies.

Excluded studies

For this update, we excluded 14 studies. Of the records identified by the original and the update search, altogether we excluded 77 studies. A number of the studies identified in the searches compared oxycodone in combination with another drug (e.g. naloxone or acetaminophen) against oxycodone alone, another active drug or placebo. Such studies were not included as they would not answer our primary question, which concerned the effectiveness of oxycodone for cancer pain. The majority of the 77 studies were excluded because they did not include the population or comparison of interest (42 studies), while others were excluded because they were systematic (11) or narrative reviews (seven), not RCTs/RCT‐based analyses (13), letters to the editor (two), case reports (one) or did not report relevant outcome data (one). See also Characteristics of excluded studies table.

Moreover, we also explored the possibility of extending our searches to cover Chinese databases. These exploratory searches in the four main Chinese databases (China Network Knowledge Infrastructure (CNKI), Chinese Scientific Journals Database (VIP), Wanfang data and SinoMed) identified 2087 de‐duplicated records, and after full‐text screening, we found over 200 potentially eligible studies that on balance we decided against including. This decision was not taken lightly but ultimately arrived at for two reasons: 1) At that point, we had already undertaken data extraction and risk of bias assessment of the 14 new eligible studies (Cao 2015; Gao 2012; Li 2013; Ren 2012; Song 2015; Sun 2013; Tu 2015; Wang 2008; Xie 2018; Ye 2012; Yu 2007; Yu 2009; Zhang 2011; Zhang 2016a) identified from the systematic review of Chinese studies (Zhou 2020), all published in Chinese and, during this work, it became apparent that all of these studies were compromised by extensive methodological under‐reporting which is illustrated by the fact that we were only able to arrive at a risk of bias assessment (low or high) in 11 out of 140 possible ratings, leaving the remaining 129 ratings as unclear (Figure 2). We had no reason to expect that this would be different for Chinese studies identified through a systematic search of Chinese databases, an assertion which is also supported by Tong 2018 and Zhang 2016b. This, in turn, would have implications for the conclusions we would be able to arrive at based on these data, and there was a risk that this extensive uncertainty would over‐shadow any conclusions we would otherwise be able to arrive at with some degree of certainty, especially given the large number of potentially relevant studies. 2) The second reason for not pursuing the search of Chinese databases concerns the understanding or employment of the term "randomised" in the context of patient allocation in Chinese studies published in Chinese journals, with a number of studies finding that a large proportion of Chinese studies described as randomised are in fact not (Tong 2018; Wu 2009a; Zhou 2019). Although the inclusion of Chinese studies identified through searches of Chinese databases would have served to limit the geographical/Western world bias of our included studies and thereby would have increased the applicability of our results, the large extent of uncertainty surrounding the methodological quality of such studies coupled with the real possibility that many of them would indeed not be randomised studies (and would therefore put the 'systematic' aspect of our review at risk due to the potential inadvertent inclusion of observational non‐randomised studies, which is an exclusion criterion of this review) meant that on balance and, in accordance with our protocol, we did not pursue the search of Chinese databases further. We note that all these concerns of course also apply to the majority of the new studies found for this update as 14 of the 18 studies were from Chinese journals, however, they were identified as part of the search strategy agreed in our protocol and were therefore included. They all examined the same comparison (CR oxycodone versus CR morphine) and we performed additional sensitivity analyses to assess the impact of their inclusion.

Figure 2

Risk of bias summary: review authors' judgements about each risk of bias item for each included study

Studies awaiting classification

Nine studies are awaiting classification because they are either ongoing but of unclear relevance, or published in such a way that not enough information is available to ascertain whether they meet the inclusion criteria (e.g. as an abstract only). In some cases, we have attempted to contact the authors, but have not successfully obtained a response. We await further information, including study completion and publication, before we can ascertain their relevance to the current review and classify them accordingly. See also Characteristics of studies awaiting classification table.

Ongoing studies

Eighteen studies are ongoing. These studies examine some new drug comparisons involving oxycodone compared to those reported in this review, and some comparisons that have already been included. Upon their completion and publication, we hope to be able to include data from all of them in future updates of this review. See also Characteristics of ongoing studies table.

Risk of bias in included studies

See Figure 2 and Figure 3 for summaries of the risk of bias judgements made for the included studies.

Figure 3

Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies

Allocation

Only seven studies were considered to be at low risk of bias for both generation of randomisation sequence and allocation concealment (Corli 2016; Heiskanen 1997; Imanaka 2013; Inoue 2018; Lux 2014; Yu 2014; Zecca 2016), while a further seven studies were considered at low risk of bias for randomisation sequence but at unclear risk of bias for allocation concealment (Cao 2015; Liu 2021; Riley 2015; Song 2015; Sun 2013; Xie 2018; Zhang 2011), and one study was at high risk of inadequate allocation concealment while at unclear risk for randomisation sequence generation (Lee 2017). The remainder of the studies did not report enough information for us to assess whether the methods employed to generate the randomisation sequence and to ensure allocation concealment were adequate. In nine studies, we could make a judgement that the treatment groups were comparable at baseline (Corli 2016; Imanaka 2013; Inoue 2017; Kaplan 1998; Liu 2021; Riley 2015; Salzman 1999; Su 2015; Zhang 2014). Three further studies reported that the groups were comparable at baseline apart from there being more participants with bone metastasis and an Eastern Cooperative Oncology Group (ECOG) performance status of 3 in the hydromorphone group than in the oxycodone group (Yu 2014), more participants with a Karnofsky Performance Status (from 0 to 100) ≤ 70 in the CR oxycodone group than in the CR morphine group (Zecca 2016), or more people with pancreatic and gastric cancers in the oxycodone group and more gastric, lung, and colorectal cancers in the morphine group (Lee 2017). In the remaining studies, it was unclear whether the participant selection methods employed had resulted in comparable, balanced groups at the start of the study.

Blinding

The problem of under‐reporting was also an issue when assigning risk of bias estimates to the items assessing performance and detection bias, that is, blinding. In only one instance was it directly and unequivocally reported who was blinded; for the most part so we had to infer, on the basis of supplementary information, whether we were reasonably certain that blinding had been adequately executed for a given individual (that is, participant, treating personnel, outcome assessors, or a combination of these, where not the participants themselves).

On this basis, the risk of performance bias was considered to be low for the primary outcome of pain in 13 of the studies (Beaver 1978a; Beaver 1978b; Bruera 1998; Hagen 1997; Heiskanen 1997; Imanaka 2013; Kaplan 1998; Lauretti 2003; Lux 2014; Mucci‐LoRusso 1998; Parris 1998; Stambaugh 2001; Yu 2014), high in nine of the studies (Corli 2016; Lee 2017; Leow 1995; Liu 2021; Mercadante 2010; Nosek 2017; Riley 2015; Salzman 1999; Zecca 2016) that were all described as open‐label (with the exceptions of Liu 2021 and Nosek 2017), and unclear in the remaining studies. For adverse events, the risk of performance bias was low in 11 studies (Beaver 1978a; Beaver 1978b; Bruera 1998; Hagen 1997; Imanaka 2013; Kaplan 1998; Lauretti 2003; Lux 2014; Mucci‐LoRusso 1998; Stambaugh 2001; Yu 2014), high in the same nine studies as was the case for pain (Corli 2016; Lee 2017; Leow 1995; Liu 2021; Mercadante 2010; Nosek 2017; Riley 2015; Salzman 1999; Zecca 2016), and unclear in the remaining studies. The pattern of judgements was identical for detection bias, for both outcomes. This was the case for the primary outcome of pain because, according to our criteria, this outcome had to be participant‐reported. It was, therefore, at risk of detection bias to the same extent that it was at risk of performance bias, since both depend on participant blinding. As is also evident from the bias judgements (see Characteristics of included studies), when a study was described as double‐blind but did not describe who was blinded, additional information in the studies generally led us to the conclusion that at least the participants seemed to be blinded, although we did not feel able to gauge with sufficient confidence who else might have been blinded. Given that it was not always clear who assessed the adverse events, this accounts for the similar judgements for performance and detection bias for this outcome.

Incomplete outcome data

Overall, the data from 88% of the total number of enrolled/randomised participants were analysed for pain and 93.1% for adverse events, which indicates that attrition bias was a substantial problem in this dataset especially for pain, with only 12 studies considered at low risk (Corli 2016; Inoue 2017; Inoue 2018; Kalso 1990; Kaplan 1998; Lee 2017; Leow 1995; Liu 2021; Parris 1998; Su 2015; Zecca 2016; Zhang 2011), and 14 studies considered at high risk (Bruera 1998; Gabrail 2004; Hagen 1997; Heiskanen 1997; Imanaka 2013; Lux 2014; Mercadante 2010; Mucci‐LoRusso 1998; Nosek 2017; Riley 2015; Salzman 1999; Stambaugh 2001; Yu 2009; Yu 2014), while the rest of the studies were at unclear risk for the primary outcome of pain. For adverse events, the risk of attrition bias was slightly less, with 16 studies considered at low risk (Corli 2016; Imanaka 2013; Inoue 2017; Inoue 2018; Kalso 1990; Kaplan 1998; Lee 2017; Leow 1995; Liu 2021; Mucci‐LoRusso 1998; Parris 1998; Salzman 1999; Su 2015; Yu 2014; Zecca 2016; Zhang 2011), eight studies considered at high risk (Bruera 1998; Hagen 1997; Heiskanen 1997; Mercadante 2010; Nosek 2017; Riley 2015; Stambaugh 2001; Yu 2009), and the remainder at unclear risk.

Selective reporting

We considered 21 of the included studies to be at low risk of selective reporting bias, whereas six of the studies either did not report adverse events ± pain or did not report them in a manner so they could be scrutinised for (and potentially included in) an evidence synthesis (Beaver 1978a; Beaver 1978b; Bruera 1998; Lauretti 2003; Lux 2014; Nosek 2017); these studies were therefore judged as being at high risk. One study only reported four adverse events in a transparent manner and was therefore considered at unclear risk of reporting bias (Hagen 1997). All of the 14 newly included Chinese studies were considered at unclear risk of selective reporting bias due to lack of information reported (see also Excluded studies).

Other potential sources of bias

Adequate titration

One study examined titration as its main objective (Salzman 1999). In the other 40 studies, the participants appeared to be adequately titrated in the majority of the studies (Bruera 1998; Corli 2016; Gabrail 2004; Hagen 1997; Heiskanen 1997; Imanaka 2013; Inoue 2017; Inoue 2018; Kalso 1990; Lauretti 2003; Lux 2014; Mucci‐LoRusso 1998; Nosek 2017; Parris 1998; Riley 2015; Stambaugh 2001; Su 2015; Yu 2009; Yu 2014; Zecca 2016; Zhang 2011; Zhang 2014), although this was not the case in one study (Kaplan 1998) and unclear in the remaining studies.

Availability of data from both time periods of cross‐over trials

For all 11 cross‐over trials, data were available for all cross‐over phases.

Other bias

Twenty‐three of the included studies were considered at low risk of any other biases (e.g. carry‐over effects in the cross‐over trials) with the remainder being judged to be at unclear risk of other bias due to the limited manner in which the trials were reported (Beaver 1978a; Beaver 1978b; Cao 2015; Gao 2012; Li 2013; Liu 2021; Ren 2012; Song 2015; Su 2015;Sun 2013; Tu 2015; Wang 2008; Xie 2018; Ye 2012; Yu 2007; Yu 2009; Yu 2014; Zhang 2011; Zhang 2014; Zhang 2016a).

Study funding

Eighteen of the included studies had received commercial funding or had authors who were employees of the drug manufacturers, or both (Beaver 1978a; Beaver 1978b; Gabrail 2004; Hagen 1997; Heiskanen 1997; Imanaka 2013; Inoue 2017; Inoue 2018; Kaplan 1998; Lee 2017; Leow 1995; Lux 2014; Mucci‐LoRusso 1998; Parris 1998; Salzman 1999; Stambaugh 2001; Yu 2014; Zecca 2016). Only seven studies were considered free from the potential influence of commercial funding (Corli 2016; Kalso 1990; Li 2013; Liu 2021; Nosek 2017; Riley 2015; Zhang 2011), with the rest of the studies having unclear status.

Effects of interventions

Analysis 1.1 and Figure 4 shows the pain intensity scores for each of the listed treatment groups, subgrouped according to overall treatment comparisons. We felt that presenting the pain intensity data this way, for the studies where it was possible, gave a comprehensive overview of the pain intensity data for the majority of the included studies, although the actual analyses should be treated with some caution as outlined in the Unit of analysis issues section. The inclusion of, in particular, the Chinese studies for this update, presented challenges in terms of analysing the pain data comprehensively as almost all of these studies only reported pain relief in categorical terms, and not pain intensity as a continuous measure. We therefore included an additional meta‐analysis of the pain data of the proportions of participants who achieved "complete" and/or "significant" pain relief (Analysis 1.2), which is illustrated in Figure 5 . Where possible, we have captured the exact definitions used by the individual studies in the Characteristics of included studies, but these were not always fully reported in the studies.

Figure 4

Forest plot of comparison: 1 Pain, outcome: 1.1 Pain intensity and pain relief (continuous)

Figure 5

Forest plot of comparison: 1 Pain, outcome: 1.2 Complete and/or significant pain relief (categorical)

Controlled‐release oxycodone versus immediate‐release oxycodone

Four studies compared CR oxycodone to IR oxycodone (Kaplan 1998; Parris 1998; Salzman 1999; Stambaugh 2001).

Pain intensity and pain relief

Pooled analysis of three of the four studies suggests that there is little to no difference in pain intensity after treatment with either CR or IR oxycodone (SMD 0.12, 95% CI ‐0.1 to 0.34; participants = 319; studies = 3; I² = 38%; Figure 4), which was also in line with the finding that none of the included studies reported that pain intensity differed between the treatment groups. Salzman 1999 could not be included in the pooled analysis due to the titration design of the study, so was instead summarised narratively below. Sensitivity analysis excluding the cross‐over trial (Stambaugh 2001) did not change the overall results although heterogeneity increased substantially (SMD 0.16, 95% CI ‐0.08 to 0.41; participants = 259; studies = 2; I² = 62%). Kaplan 1998 analysed 156 participants for efficacy evaluation rather than 160 participants; however, it was unclear from which groups these participants were missing. In the meta‐analyses, we removed two from each group, and sensitivity analyses showed none of the other possible options made any difference to the conclusions.

Salzman 1999 examined, in a parallel‐group trial lasting up to 21 days, whether CR oxycodone could be used as readily as IR oxycodone for titration to stable pain control and found that 22/24 and 19/24 participants in the CR and IR groups, respectively, achieved stable pain control within a mean time of 1.6 days (SE = 0.4) and 1.7 days (SE = 0.6), respectively.

We judged the certainty of evidence for this outcome to be very low. We downgraded the certainty of evidence by two levels for very serious limitations to study quality due to risk of bias (arising from attrition bias and under‐reporting) and by one level due to imprecision (arising from low participant numbers). See summary of findings Table 1.

Adverse events

The evidence is very uncertain about the effect of CR oxycodone compared with IR oxycodone on adverse events, including asthenia (RR 0.58, 95% CI 0.20 to 1.68; participants = 208; studies = 2; I² = 30%; Analysis 2.5), confusion (RR 0.78, 95% CI 0.20 to 3.02; participants = 157; studies = 2; I² = 25%; Analysis 2.6), constipation (RR 0.71, 95% CI 0.45 to 1.13; participants = 317; studies = 3; I² = 38%; Analysis 2.7; Figure 6), dizziness/lightheadedness (RR 0.74, 95% CI 0.40 to 1.37; participants = 317; studies = 3; I² = 15%; Analysis 2.9), drowsiness/somnolence (RR 1.03, 95% CI 0.69 to 1.54; participants = 317; studies = 3; I² = 0%; Analysis 2.10; Figure 7), dry mouth (RR 1.14, 95% CI 0.48 to 2.75; participants = 317; studies = 3; I² = 0%; Analysis 2.11), insomnia (RR 1.04, 95% CI 0.31 to 3.53; participants = 269; studies = 2; I² = 35%; Analysis 2.16), nausea (RR 0.85, 95% CI 0.56 to 1.28; participants = 317; studies = 3; I² = 0%; Analysis 2.18; Figure 8), nervousness (RR 0.57, 95% CI 0.20 to 1.64; participants = 208; studies = 2; I² = 0%; Analysis 2.20), pruritus (RR 1.46, 95% CI 0.65 to 3.25; participants = 317; studies = 3; I² = 33%; Analysis 2.21), vomiting (RR 0.66, 95% CI 0.38 to 1.15; participants = 317; studies = 3; I² = 18%; Analysis 2.23; Figure 9), and discontinuation due to adverse events (RR 0.60, 95% CI 0.29 to 1.22; participants = 317; studies = 3; I² = 0%; Analysis 2.24). The I² statistic was 55% for sweating, so the pooled results reported in Analysis 2.22 should be disregarded. The results were also very inconsistent for headache (I² statistic was 61%) and it was unclear whether there were any differences between the interventions (Analysis 2.15). Parris 1998 analysed 109 participants for safety evaluation; however, it was unclear which group had 55 and which had 54 participants. In the meta‐analyses of adverse events, we allocated 54 participants to the CR oxycodone group and 55 to the IR oxycodone group. Sensitivity analyses showed that allocating 55 participants to the CR oxycodone group and 54 participants to the IR oxycodone group made no difference to the conclusions.

Figure 6

Forest plot of comparison: 2 Adverse events, outcome: 2.7 Constipation

Figure 7

Forest plot of comparison: 2 Adverse events, outcome: 2.10 Drowsiness/somnolence

Figure 8

Forest plot of comparison: 2 Adverse events, outcome: 2.18 Nausea

Figure 9

Forest plot of comparison: 2 Adverse events, outcome: 2.23 Vomiting

In a parallel‐group trial lasting five days, Parris 1998 reported that all the adverse events observed during the study resolved. Stambaugh 2001 conducted a cross‐over study with a duration of three to seven days per phase, and stated that: "The study showed similar incidences and numbers of reports of individual adverse events considered related to the IR and CR drug" (page 505), but did not report any formal statistical comparisons of the adverse event rates between the study groups. Table 1 contains all the adverse events reported by the included studies comparing CR oxycodone and IR oxycodone.

Table 1. Controlled‐release (CR) oxycodone versus immediate‐release (IR) oxycodone: adverse events

Comparison	CR oxycodone versus IR oxycodone
Study	Kaplan 1998		Parris 1998 ^a		Salzman 1999		Stambaugh 2001
Treatment	CR	IR	CR	IR	CR	IR	CR	IR
Any adverse events	‐	‐	38/54‐55	38/54‐55	‐	‐	10/30	10/30
Total adverse events	109	186	138	142	‐	‐	‐	‐
Abdominal pain	‐	‐	3/54‐55	1/54‐55	‐	‐	‐	‐
Anxiety	0/78	4/82	‐	‐	‐	‐	‐	‐
Asthenia	3/78	8/82	‐	‐	2/24	1/24	2/30	2/30
Confusion	‐	‐	0/54‐55	2/54‐55	3/24	2/24	‐	‐
Constipation	9/78	17/82	12/54‐55	10/54‐55	4/24	9/24	1/30	1/30
Dizziness, lightheadedness	5/78	11/82	8/54‐55	10/54‐55	2/24	0/24	3/30	3/30
Drowsiness, somnolence	14/78	17/82	13/54‐55	12/54‐55	9/24	7/24	3/30	2/30
Dry mouth	3/78	5/82	4/54‐55	3/54‐55	3/24	1/24	1/30	1/30
Headache	0/78	6/82	7/54‐55	3/54‐55	1/24	1/24	‐	‐
Insomnia	2/78	4/82	3/54‐55	1/54‐55	‐	‐	‐	‐
Nausea	14/78	21/82	11/54‐55	13/54‐55	7/24	5/24	4/30	3/30
Nervousness	3/78	5/82	‐	‐	2/24	4/24	0/30	1/30
Postural hypotension	‐	‐	‐	‐	5/24	4/24	‐	‐
Pruritus	2/78	4/82	7/54‐55	5/54‐55	4/24	0/24	1/30	2/30
Sweating	4/78	3/82	1/54‐55	5/54‐55	‐	‐	2/30	1/30
Vomiting	8/78	14/82	5/54‐55	11/54‐55	5/24	3/24	2/30	0/30
Discontinuation due to adverse events	6/78	10/82	4/54‐55	7/54‐55	1/24	2/24	‐	‐

^a Total number of participants for safety evaluation = 109. Not clear which group had 55 and 54 participants, respectively.

‐: not reported
CR: controlled‐release
IR: immediate‐release

We judged the certainty of evidence for adverse events to be very low in all cases. We downgraded the certainty of the evidence by two levels due to imprecision (arising from very low event rates); and we downgraded the certainty of evidence by two levels for very serious limitations to study quality due to risk of bias (performance/detection bias, and inadequate titration and under‐reporting of the domains of selection, performance, detection and attrition bias, and whether the participants were adequately titrated).

Quality of life

There were no data for quality of life, but three studies reported treatment acceptability and their results showed that there may be little to no difference in treatment acceptability between CR and IR oxycodone. In particular, Kaplan 1998 reported in a parallel‐group study lasting six days that there was no difference in treatment acceptability between the study groups (mean at study end 3.2, SE 0.1, in both groups), and Parris 1998 found no differences in acceptability of treatment between the study groups at any time point. In Stambaugh 2001, 30/30 and 29/30 participants rated IR and CR oxycodone, respectively, as of 'fair', 'good' or 'excellent' acceptability during the last 24 hours of the treatment phases, with 24/30 and 22/30 participants rating the drugs 'good' or 'excellent', respectively. We judged the certainty of evidence for this outcome to be very low. We downgraded the certainty of evidence by two levels for very serious limitations to study quality due to risk of bias (arising from attrition bias and under‐reporting) and by one level due to imprecision (arising from low participant numbers).

Participant preference

None of the studies reported data for participant preference.

Controlled‐release oxycodone versus extended‐release oxycodone

One study compared CR oxycodone to ER oxycodone (Lux 2014).

Pain intensity and pain relief

Lux 2014 suggests there may be little to no difference in pain intensity between CR and ER oxycodone in a cross‐over trial with each of the two phases lasting 10 days (Figure 4). We judged the certainty of evidence for this outcome to be very low. We downgraded the certainty of evidence by one level for serious limitations to study quality due to risk of bias (arising from attrition bias) and by two levels due to imprecision (arising from very low participant numbers).

Adverse events

Lux 2014 also included participants with non‐malignant pain and only reported adverse events for the whole sample. Therefore, we have not reported results for adverse events.

Quality of life

The study did not report quality of life.

Participant preference

The study did not report participant preference.

Controlled‐release oxycodone versus controlled‐release morphine

Twenty‐four studies compared CR oxycodone to CR morphine (Bruera 1998; Cao 2015; Corli 2016; Gao 2012; Heiskanen 1997; Lauretti 2003; Li 2013; Mercadante 2010; Mucci‐LoRusso 1998; Nosek 2017; Ren 2012 ; Riley 2015; Song 2015, Sun 2013; Tu 2015; Wang 2008; Xie 2018; Ye 2012; Yu 2007; Yu 2009; Zecca 2016; Zhang 2011; Zhang 2014; Zhang 2016a).

Pain intensity and pain relief

Pain intensity: Fourteen studies could not be included in the pooled analysis due to the design of the study (Lauretti 2003) or because pain intensity was not reported as a continuous variable (Gao 2012; Li 2013; Nosek 2017; Ren 2012; Song 2015; Sun 2013; Tu 2015; Wang 2008; Ye 2012; Yu 2009; Zhang 2011; Zhang 2014; Zhang 2016a), and the results of these studies are therefore summarised separately below. Pooled analysis of Bruera 1998; Cao 2015; Corli 2016; Heiskanen 1997, Mercadante 2010; Mucci‐LoRusso 1998; Riley 2015; Xie 2018; Yu 2007; and Zecca 2016 showed that the pain intensity scores may be lower after treatment with CR oxycodone than with CR morphine (SMD ‐0.18, 95% CI ‐0.30 to ‐0.06; n = 1137; studies = 10; I² = 96%; Analysis 1.1; Figure 4), however, the I² indicated excessive heterogeneity and inspection of Figure 4 indicated two clear outlying studies (Cao 2015; Xie 2018) which were among the studies added from the group of Chinese language studies. Sensitivity analyses omitting the Chinese language studies (Cao 2015; Xie 2018, and Yu 2007) indicated that the pain intensity scores may be lower after treatment with CR morphine than with CR oxycodone (SMD 0.14, 95% CI 0.01 to 0.27; n = 882; studies = 7; I² = 7%). Due to the concerns outlined in Excluded studies, we consider the latter analysis our main analysis. Using the SD of the baseline average pain score of the full sample (200 participants; SD 1.94) in Riley 2015 to express this SMD as a difference in Brief Pain Inventory scores (0 to 10 numerical rating scale from no pain to worst pain imaginable) between the treatments gave an estimated difference of 0.27 between the treatments, which was not clinically significant. Moreover, sensitivity analysis excluding the two cross‐over trials (Bruera 1998, Heiskanen 1997) provided wider confidence intervals, which included no effect between CR oxycodone and CR morphine (SMD 0.12, 95% CI ‐0.02 to 0.26; n = 782; studies = 5; I² = 24%).

Pain relief: Pooled analysis of those studies that reported the number of participants who experienced complete or significant pain relief (Corli 2016; Gao 2012; Li 2013; Ren 2012; Song 2015; Sun 2013; Tu 2015; Wang 2008; Ye 2012; Yu 2007; Yu 2009; Zhang 2011; Zhang 2016a) showed that there may be little to no difference in the proportions of participants achieving complete or significant pain relief between CR oxycodone and CR morphine (RR 1.02, 95% CI 0.95 to 1.10; n = 1249; studies = 13; I² = 0%; Analysis 1.2; Figure 5). This was regardless of whether the CR oxycodone data were compared to the CR morphine sulfate or CR morphine hydrochloride data in Zhang 2016a.

Additional results reported by the individual studies: Gao 2012, in a parallel‐group trial of unclear duration, did not report pain intensity but rather that 20, 9 and 1 participants in the CR oxycodone group experienced complete, partial and no pain relief, respectively. The corresponding numbers after treatment with CR morphine were 18, 9 and 1.

In a four‐arm parallel‐group trial lasting 28 days, Corli 2016 compared CR oxycodone with CR morphine (and also included a TD fentanyl and a TD buprenorphine arm), and found that there may be little to no difference between CR oxycodone and CR morphine in terms of requirement for additional opioids (CR oxycodone: 33/125 participants; CR morphine: 36/122 participants; P = 0.59), the opioid escalation index > 5% (CR oxycodone: 24/125 participants; CR morphine: 13/122 participants; P = 0.06), or premature discontinuation for pain‐related reasons (CR oxycodone: 19/125 participants; CR morphine: 33/122 participants; P = 0.051); however, the proportion of participants requiring adjuvant drugs may be higher in the CR oxycodone group (CR oxycodone: 102/125 participants; CR morphine: 84/122 participants; P = 0.02), whereas the proportion of participants requiring switches (CR oxycodone: 15/125 participants; CR morphine: 27/122 participants; P = 0.03) may be lower in the CR oxycodone group compared to the CR morphine group.

Lauretti 2003 conducted a two‐phase (each lasting 14 days) cross‐over study to examine IR morphine consumption (which was the main outcome) during treatment with CR oxycodone and CR morphine, keeping the ratio of CR oxycodone and CR morphine constant (1:1.8). IR morphine was used as rescue medication and the participants were allowed to take as much as necessary to keep the visual analogue scale (VAS) pain score below 4. The participants consumed 38% more IR morphine during treatment with CR morphine than with CR oxycodone. Lauretti 2003 concluded that the results indicated that CR oxycodone combined with IR morphine was associated with superior analgesia and lower, or similar, rates of adverse events (see 'Adverse events' below) than a combination of CR and IR morphine.

Li 2013 conducted a parallel‐group trial lasting three days, and did not report pain intensity, but rather that 27, 13, and 2 of the 42 participants in the CR oxycodone group experienced complete, partial and mild pain relief, respectively. The corresponding numbers after treatment with CR morphine were 21, 17 and 2 out of 40 participants. Li 2013 also found that the mean onset to pain relief may be faster after CR oxycodone (mean (SD) = 44 (12.71) minutes) than after CR morphine (mean (SD) = 85 (12.96) minutes) treatment.

Mucci‐LoRusso 1998 conducted a parallel‐group trial lasting up to 12 days and found that 40/48 and 42/52 participants achieved stable pain control after receiving CR oxycodone and CR morphine, respectively, within a median of 2 days for both groups (ranges were 1 to 10 and 1 to 9 days, respectively).

Nosek 2017, in a four‐arm trial lasting 28 days comparing CR oxycodone to TD fentanyl, TD buprenorphine and CR morphine, could not be included in the analysis as the authors did not report the actual data for this outcome. Rather, they reported the pain data overall for all analysed participants, which they analysed using four treatment‐by‐time analyses of variance (ANOVA) which yielded significant interactions for all four pain variables analysed (pain at its worst; pain at its least; pain on average; pain right now). However, the authors did then not go on to perform (and present) simple main effects analyses in order to ascertain what drug and time differences underlaid these interactions. Two emails sent by our team to the authors have failed to elicit a response from the authors so, at this point, we cannot further examine these data.

In a parallel‐group trial of 14 days' duration, Ren 2012 found that out of the 40 participants in the CR oxycodone group 38, 1 and 1 achieved significant, moderate and mild pain relief, respectively. For the 40 participants in the CR morphine group, the numbers were 37, 2 and 1, respectively. Ren 2012 also found that there may be little to no difference in quality of life between the CR oxycodone (mean (SD) = 71 (8)) and CR morphine (mean (SD) = 69 (7)) groups.

In an open‐label, parallel‐group trial of one‐year duration, Riley 2015 compared CR oxycodone to CR morphine and found that 67% and 62% of the participants achieved a response to first‐line oxycodone and morphine, respectively, and their inferential analyses indicated that there may be little to no difference between these proportions. Moreover, in the participants who achieved a response to their assigned first‐line treatment, there may be little to no difference between the treatments in the five pain indices studied (that is, 'worst pain', 'least pain', 'average pain', 'pain right now', and 'percentage relief').

Song 2015 conducted a one‐month long parallel‐group trial and found that 23, 20 and 12 of the 55 participants in the CR oxycodone group experienced markedly effective, effective and ineffective pain relief, respectively. For the 55 participants who received CR morphine, the corresponding numbers were 22, 18 and 15. The mean (SD) onset to pain relief was 1.27 (0.45) hours in the CR oxycodone group and 1.59 (0.61) hours after treatment with CR morphine. Song 2015 also found that there may be little to no difference in quality of life between the CR oxycodone (mean (SD) = 37.25 (8.14)) and CR morphine (mean (SD) = 36.98 (7.59)) groups.

In a parallel‐group trial of 30 days' duration, Sun 2013 reported that, of the 102 participants in the CR oxycodone group, 81 achieved an NRS pain score of 0‐3, 20 achieved a score of 4‐6 and 1 had a score of 7‐10. In the CR oxycodone group, the corresponding numbers were 77, 24, and 1, respectively, out of a total of 102 participants. After CR oxycodone, the onset of pain relief was < 1 hour for 45 participants and > 1 hour for 57 participants. It was < 1 hour for 14 participants and > 1 hour for 88 participants in the CR morphine group.

In a parallel‐group trial by Tu 2015, which lasted > 14 days, 18, 11, 9 and 5 participants (of a total of 43 participants), respectively, achieved significant, moderate, mild and no pain relief after CR oxycodone treatment, whereas 19, 10, 8 and 6 (of a total of 43 participants), respectively, achieved significant, moderate, mild and no pain relief after CR morphine treatment.

Wang 2008 conducted a parallel‐group trial that lasted ≥ 14 days, however, the pain outcomes appear to have only been reported at 1, 2, 4, 8 and 12 hours. We have reported here those outcomes reported for 12 hours, where 9, 12, 6 and 3 (out of 30) participants, respectively, achieved complete, significant, moderate and mild pain relief in the CR oxycodone group. In the CR morphine group 10, 12, 6 and 2 (out of 30) participants, respectively, achieved complete, significant, moderate and mild pain relief. The mean (range) onset to pain relief was 43 (22‐65) minutes in the CR oxycodone group and 82 (58‐102) minutes in the CR oxycodone group. The mean (duration of?) analgesia time was 12.2 (range 8.5‐14.5) hours in the CR oxycodone group and 12.5 (range 9.5‐15.5) hours in the CR oxycodone group.

Ye 2012 in a parallel‐group trial of seven days' duration reported that 27, 11, 3 and 1 (of 42) participants, respectively, experienced complete, partial, mild and no pain relief after treatment with CR oxycodone; and that 24, 7, 6 and 4 (of 41) participants, respectively, experienced complete, partial, mild and no pain relief after treatment with CR morphine.

In a parallel‐group trial lasting five days, Yu 2007 found that on the fifth day, 6, 7 and 2 (of 15) participants experienced complete, significant and moderate pain relief, respectively, in the CR oxycodone group and 5, 7 and 3 (of 15) participants in the CR morphine group experienced complete, significant and moderate pain relief, respectively.

Yu 2009 conducted a 18‐day parallel‐group trial and found that 0, 23, 7, and 2 (of 32) participants in the CR oxycodone group achieved complete, significant, moderate and mild pain relief. The corresponding numbers were 0, 22, 6 and 2 (of 30) participants in the CR morphine group. The number of break‐through pain events were 22 and 15 in the CR oxycodone and CR morphine groups, respectively. The mean amount of rescue medication used may be lower in the CR oxycodone group (mean 23.43 (SD or SE [study did not report which] 30.23) mg) than in the CR morphine group (mean 40.33 (SD/SE 34.39) mg).

Zecca 2016, in a parallel‐group trial lasting two weeks, reported that there may be little to no difference in opioid dose escalation between treatments (CR oxycodone 8.3%; CR morphine 6.5%).

In a 4‐day long parallel‐group trial, Zhang 2011 found that 3, 21, 7 and 4 (of 35) participants, respectively, experienced complete, significant, moderate and' mild or no' pain relief in the CR oxycodone group. In the CR morphine group, the corresponding numbers of participants were 3, 19, 8 and 2 of a total of 32 participants.

Zhang 2014 conducted a three‐arm parallel‐group trial of unknown duration comparing CR oxycodone to CR morphine and CR MS Contin, and found that there may be little to no difference in pain relief rates (i.e. participants experiencing at least moderate pain relief) between the three groups (CR oxycodone 53/57 participants; CR morphine 51/57 participants; CR MS Contin 52/57 participants).

Zhang 2016a, in a three‐arm parallel‐group trial lasting one month, found that the number of participants who achieved complete, significant, moderate, mild, or no pain relief (of 40 participants in each of the 3 arms) were 9, 23, 4, 4 and 0, respectively, in the CR oxycodone group; 7, 25, 5, 3 and 0, respectively, in the CR morphine sulfate group; and 10, 24, 2, 4 and 0, respectively, in the morphine hydrochloride group.

We judged the certainty of evidence for this outcome to be low. We downgraded the certainty of evidence by two levels for very serious limitations to study quality due to risk of bias (arising from under‐reporting, performance/detection bias, and attrition bias). See summary of findings Table 2.

Adverse events

Meta‐analyses: Pooled analyses showed that for most of the adverse events the CIs were wide, including no effect as well as potential benefit and harm, for the comparison between CR oxycodone and CR morphine including anorexia/appetite loss (RR 1.20, 95% CI 0.36 to 3.94; participants = 263; studies = 3; I² = 0%; Analysis 2.4), confusion (RR 1.01, 95% CI 0.78 to 1.31; participants = 584; studies = 3; I² = 16%; Analysis 2.6), dizziness/lightheadedness (RR 0.87, 95% CI 0.58 to 1.31; participants = 941; studies = 11; I² = 0%; Analysis 2.9), drowsiness/somnolence (RR 0.88, 95% CI 0.74 to 1.05; participants = 1486; studies = 15; I² = 0%; Analysis 2.10; Figure 7), dry mouth (RR 0.97, 95% CI 0.78 to 1.22; participants = 888; studies = 5; I² = 34%; Analysis 2.11), dysuria/uroschesis (RR 0.64, 95% CI 0.38 to 1.07; participants = 887; studies = 7; I² = 0%; Analysis 2.12), nausea (RR 0.93, 95% CI 0.77 to 1.12; participants = 1388; studies = 13; I² = 0%; Analysis 2.18; Figure 8), vomiting (RR 0.81, 95% CI 0.63 to 1.04; participants = 1388; studies = 13; I² = 0%; Analysis 2.23; Figure 9), nausea and vomiting (RR 0.77, 95% CI 0.56 to 1.06; participants = 637; studies = 6; I² = 37%; Analysis 2.19), pruritus (RR 0.76, 95% CI 0.51 to 1.14; participants = 1108; studies = 8; I² = 0%; Analysis 2.21), sweating (RR 4.52, 95% CI 0.54 to 37.94; participants = 220; studies = 2; I² = 0%; Analysis 2.22), and discontinuation due to adverse events (RR 0.79, 95% CI 0.36 to 1.73; participants = 618; studies = 7; I² = 9%; Analysis 2.24). However, the RRs for constipation (RR 0.75, 95% CI 0.66 to 0.86; participants = 1894; studies = 18; I² = 38%; Analysis 2.7; Figure 6), hallucinations (RR 0.52, 95% CI 0.28 to 0.97; participants = 696; studies = 4; I² = 0%; Analysis 2.14; Figure 10) and insomnia and lethargy (RR 0.48, 95% CI 0.26 to 0.90; participants = 314; studies = 2; I² = 0%; Analysis 2.17) may be lower after treatment with CR oxycodone than after CR morphine. The I² statistic was 83% for the outcome 'any adverse events' so the pooled results reported for this outcome (Analysis 2.1) should be disregarded.

Figure 10

Forest plot of comparison: 2 Adverse events, outcome: 2.14 Hallucinations

Zhang 2014 compared CR oxycodone to both CR morphine and CR MS Contin. In the meta‐analyses of adverse events, we included CR morphine as the comparison group. Sensitivity analyses substituting the CR morphine data with the CR MS Contin data showed that whether the comparison group was CR morphine or CR MS Contin made no difference to the conclusions.

Zhang 2016a compared treatment with CR oxycodone, CR morphine sulfate and CR morphine hydrochloride and, in the analyses described above, we have used the CR morphine sulfate data. Sensitivity analyses showed that using the CR morphine hydrochloride data instead made no difference to the conclusions.

Sensitivity analysis excluding the Chinese language studies made no difference to the overall results, with the exception of constipation, which indicated that there may be little to no difference between CR oxycodone and CR morphine in constipation (RR 0.98, 95% CI 0.82 to 1.16; participants = 797; studies = 5; I² = 33%). For the outcome 'any adverse events', the I² was reduced, but still too high to pool the results (53%).

Additional results reported by the individual studies: The evidence was very uncertain about the effect of CR oxycodone versus CR morphine on adverse events. Bruera 1998 reported that: "There were no statistically significant differences by treatment in mean severity for any of the elicited adverse events or in the frequency of reporting of unelicited events" (page 3225), but presented only data on sedation and nausea VAS ratings. Corli 2016 found that there may be little to no difference between the two treatment groups in the incidence of gastralgia and breathlessness, whether they were 'any degree' or 'severe.' Severe but not 'any degree' muscle spasm myoclonus may, however, occur more often in the CR morphine group than in the CR oxycodone group. Heiskanen 1997 conducted a cross‐over trial lasting three to six days per phase and found that vomiting may be more common during morphine treatment while constipation may be more common during oxycodone treatment; and that for the remaining adverse events reported there may be little to no difference between the drugs. In a parallel‐group trial lasting four weeks (with an extension of another four weeks), Mercadante 2010 found that there may be little to no difference in the reported adverse events between the groups. Lauretti 2003 found that there may be higher rates of nausea and vomiting in the CR and IR morphine group compared to the group who had a combination of CR oxycodone and IR morphine. Nosek 2017 reported only the bowel function index data split by treatment group, which may not differ between the groups. Similarly, the number of participants discontinuing treatment due to adverse events was very low and may not differ between the groups. From the ANOVAs reported by the authors, there was an indication (by non‐significant main effects of treatment and non‐significant interactions between treatment and time) that there may be little to no difference between the groups in terms of fatigue, insomnia, drowsiness, nausea, vomiting, constipation, (loss of) appetite, dyspnoea, depression and anxiety. Riley 2015 reported that there may be little to no difference in adverse event reaction scores between oxycodone and morphine, either in first‐line responders or non‐responders. The adverse event data from the remaining studies (Cao 2015; Gao 2012; Li 2013; Ren 2012; Song 2015; Sun 2013; Tu 2015; Wang 2008; Xie 2018; Ye 2012; Yu 2007; Yu 2009; Yu 2014; Zhang 2011; Zhang 2016a) have all been included in the meta‐analyses reported above with the following exceptions: mental disorders (Song 2015), dizziness and fatigue (Wang 2008), abdominal distension (Xie 2018) and insanity and 'other adverse events' (Sun 2013). They can all be found in Table 2 and Table 3 which contain all the adverse events reported by the included studies comparing CR oxycodone and CR morphine.

Table 2. Controlled‐release (CR) oxycodone versus CR morphine (English‐language studies): adverse events

Comparison	CR oxycodone versus CR morphine
Study	Bruera 1998		Corli 2016		Heiskanen 1997		Lauretti 2003		Mercadante 2010 ^a		Mucci‐LoRusso 1998		Nosek 2017		Riley 2014		Zhang 2014			Zecca 2016
Treatment	Oxy	Mor	Oxy	Mor	Oxy	Mor	Oxy	Mor	Oxy	Mor	Oxy	Mor	Oxy	Mor	Oxy	Mor	Oxy	Mor	MS Contin	Oxy	Mor
Any adverse events	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	40/48	39/52	‐	‐	‐	‐	25/57	34/57	31/57	77/81	79/94
Severe/moderate adverse events	‐	‐	48.8%	58.9%	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐
Abnormal dreams	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	3/81	1/72	‐	‐	‐	‐	‐
Anorexia, appetite loss	‐	‐	‐	‐	0/27	1/27	14/22	13/22	‐	‐	‐	‐	‐	‐	1/81	0/72	‐	‐	‐	‐	‐
Chills	‐	‐	‐	‐	1/27	0/27	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐
Confusion ‐ serious	‐	‐	55/129 12/129	59/129 20/129	‐	‐	‐	‐	0.37 (0.49)	0.25 (0.44)	‐	‐	‐	‐	7/81 3/81	2/72 0/72	‐	‐	‐	12/85	12/88
Constipation ‐ serious	‐	‐	75/129 40/129	82/129 50/129	18/27	14/27	4/22	5/22	0.63 (0.68)	0.7 (0.92)	10/48	10/52	1.91 (1.15)^b	2.14 (1.01)^b	18/81 2/81	24/72 5/72	6/57	3/57	5/57	30/85	22/87
Decreased mobility	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	0/81	2/72	‐	‐	‐	‐	‐
Depression	‐	‐	‐	‐	1/27	0/27	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐
Diarrhoea	‐	‐	‐	‐	2/27	2/27	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐
Dizziness, lightheadedness	‐	‐	‐	‐	6/27	6/27	‐	‐	‐	‐	4/48	7/52	‐	‐	3/81	2/72	2/57	3/57	4/57	‐	‐
Double vision	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	0/81	1/72	‐	‐	‐	‐	‐
Drowsiness, somnolence ‐ serious (with/without hallucinations)	‐	‐	74/129 34/129	79/129 38/129	‐	‐	7/22	11/22	0.37 (0.6)	0.35 (0.59)	7/48	10/52	‐	‐	12/81 1/81	13/72 0/72	1/57	1/57	1/57	27/85	31/88
Drunken feeling	‐	‐	‐	‐	1/27	1/27	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐
Dry mouth ‐ severe	‐	‐	66/129 27/129	66/129 31/129	12/27	15/27	3/22	2/22	0.63 (0.68)	0.6 (0.68)	1/48	7/52	‐	‐	3/81	2/72	‐	‐	‐	19/85	14/88
Dyspnoea ‐severe	‐	‐	12/129 3/129	17/129 4/129	2/27	2/27	0/22	0/22	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐
Dysuria ‐ severe	‐	‐	17/129 4/129	22/129 2/129	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	0/57	2/57	1/57	‐	‐
Extrasystoles	‐	‐	‐	‐	1/27	0/27	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐
Faecal incontinence	‐	‐	‐	‐	1/27	1/27	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐
Fall	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	0/81	3/72	‐	‐	‐	‐	‐
Feeling abnormal	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	0/81	1/72	‐	‐	‐	‐	‐
Flatus	‐	‐	‐	‐	0/27	1/27	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐
Gastralgia ‐ severe	‐	‐	21/129 6/129	24/129 3/129	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐
Hallucinations ‐ severe	‐	‐	8/129 1/129	17/129 6/129	‐	‐	0/22	0/22	‐	‐	0/48	2/52	‐	‐	3/81	4/72	‐	‐	‐	3/91	4/94
Hollow feeling	‐	‐	‐	‐	1/27	0/27	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐
Lethargy	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	1/81	0/72	‐	‐	‐	‐	‐
Memory impairment	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	1/81	1/72	‐	‐	‐	‐	‐
Muscle twitches	‐	‐	‐	‐	1/27	1/27	‐	‐	‐	‐	‐	‐	‐	‐	0/81	2/72	‐	‐	‐	‐	‐
Muscle spasm myoclonus ‐ severe	‐	‐	23/129 0/129	14/129 6/129	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐
Nausea ‐ serious (with vomiting)/severe	12.3	13.9	63/129 22/129	64/129 19/129	14/27	16/27	1/22	8/22	0.84 (0.9)	0.6 (0.75)	6/48	8/52	‐	‐	10/81 1/81	6/72 0/72	11/57	15/57	14/57	18/85	13/88
Nightmares	‐	‐	‐	‐	0/27	3/27	‐	‐	‐	‐	‐	‐	‐	‐	2/81	0/72	‐	‐	‐	‐
Pain	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	0/81	1/72	‐	‐	‐	‐	‐
Paresthesia	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	1/81	0/72	‐	‐	‐	‐	‐
Pruritus ‐ severe	‐	‐	16/129 2/129	24/129 3/129	10/27	7/27	1/22	1/22	‐	‐	4/48	5/52	‐	‐	3/81	2/72	‐	‐	‐	6/85	5/88
Sedation	21.4	25	‐	‐	16/27	18/27	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐
Sensation of empty head	‐	‐	‐	‐	‐	‐	1/22	0/11	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐
Slow speech	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	1/81	0/72	‐	‐	‐	‐	‐
Sweating, hyperhidrosis	‐	‐	‐	‐	12/27	9/27	‐	‐	‐	‐	‐	‐	‐	‐	2/81	0/72	‐	‐	‐	‐	‐
Serious toxicity secondary to infection	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	1/81	0/72	‐	‐	‐	‐	‐
Urinary hesitation	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	0/81	1/72	‐	‐	‐	‐	‐
Visual impairment	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	1/81	0/72	‐	‐	‐	‐	‐
Vomiting ‐ severe	‐	‐	29/129 12/129	35/129 12/129	5/27	10/27	0/22	7/22	‐	‐	6/48	5/52	‐	‐	9/81	4/72	5/57	10/57	6/57	11/85	10/88
Discontinuation due to adverse events	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	3/48	6/52	0/16	1/14	‐	‐	0/57	0/57	0/57	6/91	3/94
Unexpected serious adverse events	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	2/81	7/72	‐	‐	‐	‐	‐

^a Mean (SD) ratings (out of 3) experienced during week 4

^bBowel function index mean (standard deviation) (range 0‐10, with constipation indicated if score = 3 or above) at 28 days

‐: not reported
CR: controlled‐release
Mor: morphine
Oxy: oxycodone

Table 3. Controlled‐release (CR) oxycodone versus CR morphine (Chinese‐language studies): adverse events

Comparison	CR oxycodone versus CR morphine
Study	Treat‐ ment	Adverse event (AE)
Study	Treat‐ ment	Any AE	Se‐ ve‐ re AE	Ab‐ domi‐ nal dis‐ ten‐ sion	Ano‐ rexia/ appe‐ tite loss	Con‐ sti‐ pa‐ tion	Dizzi‐ ness	Dizzi‐ ness & fa‐ tigue	Drow‐ si‐ ness	Drow‐ siness & fa‐ tigue	Dry mouth	Dys‐ uria/ uro‐ sche‐ sis	Insa‐ nity/ men‐ tal disor‐ ders	In‐ som‐ nia & le‐ thar‐ gy	Nau‐ sea	Nau‐ sea & vo‐ mi‐ ting	Pru‐ ritis	Swea‐ ting	Res‐ pira‐ tory de‐ pres‐ sion	Vo‐ mi‐ ting	Dis‐ con‐ tinu‐ ation due to AE
Cao 2015	Oxy	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	3/65	‐	‐	‐	‐	1/65	‐
Cao 2015	Mor	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	11/65	‐	‐	‐	‐	7/65	‐
Gao 2012	Oxy	‐	‐	‐	‐	2/30	3/30	‐	1/30	‐	‐	‐	‐	‐	4/30	‐	‐	‐	‐	2/30	‐
Gao 2012	Mor	‐	‐	‐	‐	5/28	2/28	‐	2/28	‐	‐	‐	‐	‐	3/28	‐	‐	‐	‐	3/28	‐
Li 2013	Oxy	16/42	‐	‐	‐	9/42	‐	‐	4/42	‐	‐	0/42	‐	‐	3/42	‐	‐	‐	‐	0/42	‐
Li 2013	Mor	24/40	‐	‐	‐	12/40	‐	‐	5/40	‐	‐	1/40	‐	‐	4/40	‐	‐	‐	‐	2/40	‐
Ren 2012	Oxy	‐	0/40	‐	2/40	12/40	7/40	‐	5/40	‐	‐	‐	‐	‐	9/40	‐	2/40	‐	‐	13/40	‐
Ren 2012	Mor	‐	0/40	‐	1/40	22/40	6/40	‐	6/40	‐	‐	‐	‐	‐	8/40	‐	2/40	‐	‐	14/40	‐
Song 2015	Oxy	‐	‐	‐	‐	5/55	‐	‐	‐	‐	‐	‐	1/55	3/55	‐	6/55	2/55	‐	‐	‐	‐
Song 2015	Mor	‐	‐	‐	‐	13/55	‐	‐	‐	‐	‐	‐	7/55	10/55	‐	15/55	4/55	‐	‐	‐	‐
Sun 2013 ^a	Oxy	64/102	‐	‐	‐	19/102	‐	‐	‐	‐	1/102	0/102	2/102	10/102	‐	25/102	3/102	‐	‐	‐	‐
Sun 2013 ^a	Mor	99/102	‐	‐	‐	36/102	‐	‐	‐	‐	4/102	3/102	6/102	17/102	‐	21/102	5/102	‐	‐	‐	‐
Tu 2015	Oxy	‐	‐	‐	‐	5/43	3/43	‐	‐	2/43	‐	1/43	‐	‐	‐	5/43	‐	‐	‐	‐	‐
Tu 2015	Mor	‐	‐	‐	‐	12/43	3/43	‐	‐	3/43	‐	1/43	‐	‐	‐	4/43	‐	‐	‐	‐	‐
Wang 2008	Oxy	‐	‐	‐	‐	9/30	‐	1/30	2/30	‐	‐	1/30	‐	‐	4/30	‐	‐	‐	0/30	3/30	0/30
Wang 2008	Mor	‐	‐	‐	‐	18/30	‐	2/30	4/30	‐	‐	2/30	‐	‐	5/30	‐	‐	‐	0/30	4/30	0/30
Xie 2018	Oxy	3/48	‐	0/48	‐	1/48	1/48	‐	‐	‐	‐	‐	‐	‐	‐	1/48	‐	‐	‐	‐	‐
Xie 2018	Mor	10/47	‐	2/47	‐	2/47	3/47	‐	‐	‐	‐	‐	‐	‐	‐	3/47	‐	‐	‐	‐	‐
Ye 2012	Oxy	18/42	‐	‐	‐	3/42	2/42	‐	2/42	‐	‐	1/42	‐	‐	7/42	‐	‐	‐	‐	3/42	‐
Ye 2012	Mor	28/41	‐	‐	‐	5/41	2/41	‐	3/41	‐	‐	1/41	‐	‐	10/41	‐	‐	‐	‐	7/41	‐
Yu 2007	Oxy	‐	‐	‐	2/15	5/15	5/15	‐	2/15	‐	‐	‐	‐	‐	1/15	‐	2/15	‐	‐	2/15	‐
Yu 2007	Mor	‐	‐	‐	3/15	7/15	4/15	‐	2/15	‐	‐	‐	‐	‐	2/15	‐	3/15	‐	‐	3/15	‐
Yu 2009	Oxy	‐	‐	‐	‐	7/32	8/32	‐	5/32	‐	‐	‐	‐	‐	‐	10/32	‐	‐	‐	‐	1/32
Yu 2009	Mor	‐	‐	‐	‐	15/30	7/30	‐	4/30	‐	‐	‐	‐	‐	‐	18/30	‐	‐	‐	‐	3/30
Zhang 2011	Oxy	‐	‐	‐	‐	15/35	‐	‐	3/35	‐	‐	‐	‐	‐	10/35	‐	‐	2/35	‐	4/35	0/35
Zhang 2011	Mor	‐	‐	‐	‐	13/32	‐	‐	1/32	‐	‐	‐	‐	‐	10/32	‐	‐	0/32	‐	4/32	0/32
Zhang 2016a	Oxy	6/40	‐	‐	‐	0/40	1/40	‐	2/40	‐	‐	‐	‐	‐	‐	3/40	‐	‐	‐	‐	‐
	Mor sulfate	14/40	‐	‐	‐	2/40	5/40	‐	4/40	‐	‐	‐	‐	‐	‐	3/40	‐	‐	‐	‐	‐
	Mor hydro‐ chloride	13/40	‐	‐	‐	1/40	4/40	‐	5/40	‐	‐	‐	‐	‐	‐	3/40	‐		‐	‐	‐

^aSun 2013 also reported that 4/102 participants in the CR oxycodone group and 8/102 participants in the CR morphine group experienced "Other adverse events".

‐: not reported
AE: adverse event
CR: controlled‐release
Mor: morphine
Oxy: oxycodone

We judged the certainty of evidence for this outcome to be low or very low in all cases. We downgraded the certainty of evidence by two levels for very serious limitations to study quality due to risk of bias (arising from under‐reporting, performance/detection bias, attrition bias, or a combination of these), and we downgraded the certainty of the evidence by no, one or two levels due to imprecision (arising from low or very low event rates).

Quality of life

The evidence was very uncertain about the effect of CR oxycodone versus CR morphine on quality of life. Mucci‐LoRusso 1998 reported no clinically significant changes in quality of life for either treatment group, but did not show results or analyses. The authors also found there may be little to no difference in treatment acceptability between the treatment groups with 74% and 77% of the CR oxycodone and CR morphine participants, respectively, rating the acceptability of treatment as good to excellent and the mean acceptability ratings at the study end being 4 (SE = 0.1) in the CR oxycodone and 3.9 (SE = 0.1) in the CR morphine participants. Heiskanen 1997 found that the mean daily acceptability of treatment ratings may be higher for morphine (3.49/5; SE = 0.12) than for oxycodone (3.19/5; SE = 0.11), but Lauretti 2003 also found that there may be little to no difference in treatment acceptance between treatment with CR and IR morphine and treatment with the combination of CR oxycodone and IR morphine. Again, Nosek 2017 did not report the quality of life results by treatment group or analyse any significant interactions fully, which means that it is impossible to know what the results were. We are, therefore, limited to reporting what the authors reported, which was that there were no significant differences between the groups in emotional functioning or overall quality of life, however, there was a significant interaction between treatment and time for "well‐being" (not further analysed) and a significant main effect for physical functioning, which the authors reported as "the most beneficial effect was obtained for morphine" (page 3217). Finally, the Karnofsky Performance Scale score, which was reported by treatment, may not differ between the (four) treatment groups (CR oxycodone mean = 65, SD = 11.55; CR morphine mean = 66.67, SD = 9.85).

We judged the certainty of evidence for this outcome to be very low. We downgraded the certainty of evidence by two levels for very serious limitations to study quality due to risk of bias (arising from under‐reporting and performance, detection and attrition bias), and we downgraded one level for imprecision (arising from low participant numbers).

Participant preference

The evidence was very uncertain about the effect of CR oxycodone versus CR morphine on participant preference. In a cross‐over trial with each phase lasting seven days, Bruera 1998 reported that 8/23 participants preferred CR oxycodone treatment while 11/23 participants preferred treatment with CR morphine. We judged the certainty of evidence for this outcome to be very low. We downgraded the certainty of evidence by two levels for very serious limitations to study quality due to risk of bias (arising from under‐reporting and attrition bias), and we downgraded the certainty of the evidence by two levels due to imprecision (arising from very low participant numbers).

Controlled‐release oxycodone versus controlled‐release hydromorphone

One study compared CR oxycodone to CR hydromorphone (Hagen 1997).

Pain intensity and pain relief

The evidence was very uncertain about the effect of CR oxycodone versus CR hydromorphone on pain intensity. In a cross‐over trial lasting seven days per phase, Hagen 1997 found that there may be little to no difference in pain intensity between treatment with CR oxycodone and CR hydromorphone (Figure 4). We judged the certainty of evidence for this outcome to be very low. We downgraded the certainty of evidence by two levels for very serious limitations to study quality due to risk of bias (arising from under‐reporting and attrition bias), and two levels for imprecision (arising from very low participant numbers).

Adverse events

The evidence was very uncertain about the effect of CR oxycodone versus CR hydromorphone on adverse events. Hagen 1997 observed that there may be little to no difference in the frequency of adverse events between treatment groups with the exception of drowsiness, which may occur more frequently during treatment with oxycodone (see Table 4). We judged the certainty of evidence for this outcome to be very low. We downgraded the certainty of evidence by two levels for very serious limitations to study quality due to risk of bias (arising from under‐reporting and attrition bias), and two levels for imprecision (arising from very low participant/event numbers).

Table 4. Other oral oxycodone comparisons: adverse events

Comparison	CR oxycodone versus CR hydromorphone		CR oxycodone versus ER hydromorphone				CR oxycodone versus ER oxymorphone		CR oxycodone versus ER tapentadol		CR oxycodone versus oral ibuprofen		IR oxycodone versus IR hydromorphone
Study	Hagen 1997		Inoue 2017		Yu 2014		Gabrail 2004		Imanaka 2013		Liu 2021		Inoue 2018
Treatment	Oxy	Hyd	Oxy	Hyd	Oxy	Hyd	Oxy	Oxymo	Oxy	Tap	Oxy	Ibu	Oxy	Hyd
Any adverse events	‐	‐	77/92	71/88	117/126	111/128	‐	‐	155/172	147/168	‐	‐	65/84	73/88
Total adverse events	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐
Severe adverse events	‐	‐	10/92	7/88	30/126	30/128	‐	‐	‐	‐	0/34	0/32	‐	‐
Serious adverse events (including death)	‐	‐	14/92	11/88	18/126	11/128	‐	‐	‐	‐	0/34	0/32	8/84	7/88
Serious adverse events (including death), considered related to study drugs	‐	‐	6/92	4/88	‐	‐	‐	‐	‐	‐	0/34	0/32	3/84	4/88
Abdominal discomfort	‐	‐	‐	‐	7/126	4/128	‐	‐	‐	‐	5/34	10/32	‐	‐
Abdominal distension	‐	‐	‐	‐	7/126	7/128	‐	‐	‐	‐	‐	‐	‐	‐
Anaemia	‐	‐	‐	‐	14/126	14/128	‐	‐	‐	‐	‐	‐	‐	‐
Anorexia, appetite loss	‐	‐	3/92	7/88	21/126	20/128	‐	‐	24/172	23/168	‐	‐	‐	‐
Asthenia	‐	‐	‐	‐	9/126	11/128	‐	‐	‐	‐	‐	‐	‐	‐
Bone marrow failure	‐	‐	‐	‐	9/126	9/128	‐	‐	‐	‐	‐	‐	‐	‐
Chest discomfort	‐	‐	‐	‐	6/126	9/128	‐	‐	‐	‐	‐	‐	‐	‐
Confusion	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐
Constipation	‐	‐	14/92	11/88	45/126	43/128	19/41	21/43	64/172	51/168	9/34	3/32	19/84	21/88
Delirium	‐	‐	‐	‐	‐	‐	‐	‐	6/172	10/168	‐	‐	5/84	2/88
Diarrhoea	‐	‐	17/92	16/88	9/126	12/128	‐	‐	19/172	11/168	‐	‐	9/84	12/88
Dizziness or lightheadedness	‐	‐	5/92	6/88	22/126	21/128	9/41	7/43	‐	‐	10/34	4/32	‐	‐
Drowsiness, somnolence	28/31	19/31	18/92	23/88	‐	‐	‐	‐	36/172	29/168	3/34	0/32	21/84	23/88
Fever	‐	‐	5/92	7/88	27/126	24/128	‐	‐	‐	‐	‐	‐	‐	‐
Hallucinations	0/31	2/31	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐
Hypoproteinaemia	‐	‐	‐	‐	5/126	9/128	‐	‐	‐	‐	‐	‐	‐	‐
Insomnia	‐	‐	‐	‐	‐	‐	‐	‐	11/172	9/168	‐	‐	‐	‐
Malaise	‐	‐	6/92	3/88	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐
Nausea	15 (3)^a	13 (3)^a	21/92	36/88	45/126	43/128	15/41	17/43	61/172	48/168	3/34	0/32	14/84	14/88
Neutrophil count decreased	‐	‐	‐	‐	5/126	7/128	‐	‐	‐	‐	‐	‐	‐	‐
Oedema, peripheral	‐	‐	‐	‐	6/126	11/128	‐	‐	‐	‐	‐	‐	‐	‐
Platelet count decreased	‐	‐	‐	‐	7/126	8/128	‐	‐	‐	‐	‐	‐	‐	‐
Pruritus	‐	‐	‐	‐	‐	‐	8/41	13/43	‐	‐	‐	‐	‐	‐
Rash	‐	‐	‐	‐	4/126	7/128	‐	‐	‐	‐	‐	‐	‐	‐
Respiratory depression	‐	‐	0/92	0/88	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐
Sedation	24 (4)^a	18 (3)^a	‐	‐	‐	‐	13/41	18/43	‐	‐	‐	‐	‐	‐
Sweating or hyperhidrosis	‐	‐	‐	‐	8/126	3/128	9/41	12/43	‐	‐	‐	‐	‐	‐
Transaminases elevated	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	2/34	7/32	‐	‐
Urinary retention	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐
Urinary tract infection	‐	‐	‐	‐	7/126	4/128	‐	‐	‐	‐	‐	‐	‐	‐
Urination difficulty	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	1/34	0/32	‐	‐
Vomiting	‐	‐	16/92	32/88	47/126	43/128	7/41	5/43	41/172	42/168	‐	‐	15/84	17/88
White blood cell count decreased	‐	‐	‐	‐	17/126	13/128	‐	‐	‐	‐	‐	‐	‐	‐
Discontinuation due to adverse events	‐	‐	14/92	10/88	18/126	19/128	‐	‐	29/172	22/168	‐	‐	8/84	6/88
Deaths due to study drug treatment	‐	‐	‐	‐	0/126	0/128	‐	‐	‐	‐	0/34	0/32	‐	‐

^aMean (standard error) visual analogue scale across all days

‐: not reported
CR: controlled‐release
ER: extended‐release
Hyd: hydromorphone
Ibu: Ibuprofen
IR: Immediate‐release
IV: intravenous
Mor: morphine
Oxy: oxycodone
Oxymo: oxymorphone
Tap: tapentadol

Quality of life

The study did not report quality of life.

Participant preference

The evidence was very uncertain about the effect of CR oxycodone versus CR hydromorphone on participant preference. Hagen 1997 found that 25.8% of participants had no treatment preference, with approximately half of the remaining participants preferring oxycodone (35.5%) while the other half preferred hydromorphone (38.7%). We judged the certainty of evidence for this outcome to be very low. We downgraded the certainty of evidence by two levels for very serious limitations to study quality due to risk of bias (arising from under‐reporting and attrition bias), and two levels for imprecision (arising from very low participant numbers).

Controlled‐release oxycodone versus extended‐release hydromorphone

Two studies compared CR oxycodone to ER hydromorphone (Inoue 2017; Yu 2014).

Pain intensity and pain relief

The evidence was very uncertain about the effect of CR oxycodone versus ER hydromorphone on pain intensity and pain relief. Pooled analysis including both the parallel‐group Inoue 2017 trial of seven days' duration and the 28‐day parallel‐group trial by Yu 2014 showed that there may be little to no difference in pain intensity between the two treatment groups (Analysis 1.1; Figure 4). Yu 2014 also found that there may be little to no difference in pain intensity or pain relief between treatment with CR oxycodone and ER hydromorphone on any of the following additional pain measures: mean pain on average (CR oxycodone 3.3; ER hydromorphone 2.9), mean pain at its least in past 24 hours (CR oxycodone 1.9; ER hydromorphone 1.6), mean pain 'right now' (CR oxycodone 2.8; ER hydromorphone 2.7), mean pain relief in past 24 hours (CR oxycodone 62.2%; ER hydromorphone 64.5%), and the number of rescue medication doses taken during the overall maintenance phase (CR oxycodone 29.3; ER hydromorphone 24.2). We judged the certainty of evidence for this outcome to be very low. We downgraded the certainty of evidence by two levels for very serious limitations to study quality due to risk of bias (arising from under‐reporting of selection, performance and detection bias, and attrition bias), and one level for imprecision (arising from low participant numbers).

Adverse events

The evidence was very uncertain about the effect of CR oxycodone versus ER hydromorphone on adverse events. Pooled analysis of the event rates showed that there may be little to no difference between the treatment groups for any adverse event (RR 1.06, 95% CI 0.98 to 1.14; participants = 434; studies = 2; I² = 0%; Analysis 2.1), severe adverse events (RR 1.08, 95% CI 0.73 to 1.62; participants = 434; studies = 2; I² = 0%; Analysis 2.2), serious adverse events including death (RR 1.44, 95% CI 0.86 to 2.39; participants = 434; studies = 2; I² = 0%; Analysis 2.3), appetite loss/anorexia (RR 0.89, 95% CI 0.54 to 1.49; participants = 434; studies = 2; I² = 42%; Analysis 2.4), constipation (RR 1.10, 95% CI 0.80 to 1.49; participants = 434; studies = 2; I² = 0%; Analysis 2.7; Figure 6), diarrhoea (RR 0.91, 95% CI 0.55 to 1.49; participants = 434; studies = 2; I² = 0%; Analysis 2.8), dizziness (RR 1.00, 95% CI 0.61 to 1.64; participants = 434; studies = 2; I² = 0%; Analysis 2.9), fever (RR 1.04, 95% CI 0.66 to 1.62; participants = 434; studies = 2; I² = 0%; Analysis 2.13) and discontinuation of treatment due to adverse events (RR 1.09, 95% CI 0.69 to 1.75; participants = 434; studies = 2; I² = 0%; Analysis 2.24). For both nausea and vomiting, the I² statistics were very high (80% and 86%, respectively), so the pooled results in Figure 8 (Analysis 2.18) and Figure 9 (Analysis 2.23), respectively, for this treatment subgroup should be ignored. For both nausea and vomiting, Inoue 2017 reported that the rates may be lower in the CR oxycodone group compared to the ER hydromorphone group, whereas Yu 2014 found that there may be little to no difference between the interventions. See also Table 4 for other adverse events that were reported only by one or the other of the studies. We judged the certainty of evidence for this outcome to be moderate (for 'any adverse event') or very low (for all other reported adverse events), depending on the event rates of the individual adverse events. We downgraded the certainty of evidence by one level for risk of bias (due to under‐reporting of selection, performance and detection bias) and by none or two levels for imprecision (arising from very low event rates).

Quality of life

The studies did not report quality of life.

Participant preference

The studies did not report participant preference.

Controlled‐release oxycodone versus extended‐release oxymorphone

One study compared CR oxycodone to ER oxymorphone (Gabrail 2004).

Pain intensity and pain relief

The evidence was very uncertain about the effect of CR oxycodone versus ER oxymorphone on pain intensity. Gabrail 2004, in a cross‐over trial with each phase lasting seven to 10 days, found that there may be little to no difference in mean 24‐hour average daily pain intensity ratings between the two treatments (Analysis 1.1; Figure 4). We judged the certainty of evidence for this outcome to be very low. We downgraded the certainty of evidence by two levels for very serious limitations to study quality due to risk of bias (arising from under‐reporting and attrition bias), and by two levels for imprecision (arising from very low participant numbers).

Adverse events

The evidence was very uncertain about the effect of CR oxycodone versus ER oxymorphone on adverse events. Gabrail 2004 reported that there may be little to no difference in adverse event rates between the drugs (see Table 4), and no participants withdrew due to abnormal laboratory values, insufficient analgesia, or loss to follow‐up. We judged the certainty of evidence for this outcome to be very low. We downgraded the certainty of evidence by two levels for very serious limitations to study quality due to risk of bias (arising from under‐reporting of the domains of selection, performance, detection, and attrition bias), and by two levels for imprecision (arising from very low participant numbers).

Quality of life

The evidence was very uncertain about the effect of CR oxycodone versus ER oxymorphone on quality of life. Gabrail 2004 reported that there may be little to no difference in quality of life (general activity, mood, walking ability, normal work, relationships with others, sleep, and enjoyment of life) between the drugs. We judged the certainty of evidence for this outcome to be very low. We downgraded the certainty of evidence by two levels for very serious limitations to study quality due to risk of bias (arising from under‐reporting of the domains of selection, performance, detection, and attrition bias), and by two levels for imprecision (arising from very low participant numbers).

Participant preference

The evidence was very uncertain about the effect of CR oxycodone versus ER oxymorphone on participant preference. The study reported no data for participant preference, but Gabrail 2004 found that 78.3% of participants rated oxycodone as 'excellent,' 'very good,' or 'good' with 86.4% of the participants giving oxymorphone such ratings. We judged the certainty of evidence for this outcome to be very low. We downgraded the certainty of evidence by two levels for very serious limitations to study quality due to risk of bias (arising from under‐reporting of the domains of selection, performance, and detection bias), and by two levels for imprecision (arising from very low participant numbers).

Controlled‐release oxycodone versus extended‐release tapentadol

One study compared CR oxycodone to ER tapentadol (Imanaka 2013).

Pain intensity and pain relief

The evidence was very uncertain about the effect of CR oxycodone versus ER tapentadol on pain intensity. Imanaka 2013, in a parallel‐group trial of four weeks' duration, found that there may be little to no difference in pain intensity between the study groups (see Figure 4) with 82/139 CR oxycodone participants and 80/126 ER tapentadol participants reporting ≥ 30% improvement in pain intensity during the last three days of treatment, and 59/139 CR oxycodone participants and ER 63/126 tapentadol participants reporting ≥ 50% improvement in pain intensity during the last three days of treatment (Figure 5). We judged the certainty of evidence for this outcome to be low. We downgraded the certainty of evidence by one level for serious limitations to study quality due to risk of bias (arising from attrition bias), and by one level for imprecision (arising from low participant numbers).

Adverse events

The evidence was very uncertain about the effect of CR oxycodone versus ER tapentadol on adverse events. Inspection of Table 4 suggested that there may be little to no difference in adverse event rates between the treatment groups, but Imanaka 2013 did not present any formal statistical analyses of this apparent equality. We judged the certainty of evidence for this outcome to be high (for 'any adverse events') or low (all other reported adverse events), depending on the event rates of the individual adverse events. We downgraded the certainty of evidence by no or two levels for imprecision (arising from very low event rates).

Quality of life

The study did not report quality of life.

Participant preference

The study did not report participant preference.

Controlled‐release oxycodone versus transdermal fentanyl

Three studies compared CR oxycodone to TD fentanyl (Corli 2016; Nosek 2017; Su 2015).

Pain intensity and pain relief

The evidence was very uncertain about the effect of CR oxycodone versus TD fentanyl on pain intensity and pain relief. Pooled analysis including Corli 2016 and Su 2015 showed that there may be little to no difference in pain intensity scores after treatment with CR oxycodone or TD fentanyl (SMD 0.02, 95% CI ‐0.19 to 0.24; participants = 329; studies = 2; I² = 0%; Analysis 1.1; Figure 4). This was also the case when pain relief was analysed as a dichotomous measure (RR 0.98, 95% CI 0.85 to 1.14; participants = 329; studies = 2; I² = 0%; Analysis 1.2; Figure 5). For Nosek 2017, the situation was the same for this comparison as for the comparison between CR oxycodone and CR morphine reported above.

Corli 2016 compared CR oxycodone with TD fentanyl in a four‐arm trial of 28 days' duration, which also included a TD buprenorphine and a CR morphine group, and found that there may be little to no difference between CR oxycodone and TD fentanyl treatment in terms of the requirement for additional opioids (CR oxycodone 33/125 participants; TD fentanyl 46/124 participants), premature discontinuations for pain‐related reasons (CR oxycodone 19/125 participants; TD fentanyl 18/124 participants), proportion of participants requiring adjuvant drugs (CR oxycodone 102/125 participants; TD fentanyl 100/124 participants), or proportion of participants requiring switches (CR oxycodone 15/125 participants; TD fentanyl 16/124 participants). However, a higher proportion of participants in the CR fentanyl group may have required a mean increase in the opioid daily dose > 5% according to the opioid escalation index compared to the oxycodone group (CR oxycodone 24/125 participants; TD fentanyl 45/124 participants). In a parallel‐group trial lasting two weeks, Su 2015 found that there may be little to no difference between groups treated with CR oxycodone or TD fentanyl in pain relief (CR oxycodone 90.48%; TD fentanyl 92.11%) or pain intensity measured by response categories (CR oxycodone: 15 complete response, 21 partial response, 2 minor response, 4 no response; TD fentanyl: 13 complete response, 20 partial response, 2 minor response, 3 no response).

We judged the certainty of evidence for this outcome to be very low. We downgraded the certainty of evidence by two levels for very serious limitations to study quality due to risk of bias (arising from under‐reporting, performance bias, detection bias, attrition bias and selective reporting bias), and by one level for imprecision (arising from low participant/event numbers).

Adverse events

The evidence was very uncertain about the effect of CR oxycodone versus TD fentanyl on adverse events. Pooled analysis of the event rates for constipation and dysuria revealed that the I² statistic was 83% for constipation, so the pooled results shown in Figure 6 (Analysis 2.7) for this treatment subgroup should be ignored, and that there may be little to no difference between treatment groups for dysuria (RR 1.15, 95% CI 0.62 to 2.16; participants = 336; studies = 2; I² = 0%; Analysis 2.12).

Corli 2016 found that there may be little to no difference between the two treatment groups in the incidence of the following adverse events, whether they were 'any degree' or 'severe': drowsiness, confusion, nausea, vomiting, constipation, dry mouth, hallucinations, gastralgia, muscle spasm myoclonus, breathlessness, and itching. For Nosek 2017, the situation was the same for this comparison as for the comparison between CR oxycodone and CR morphine reported above. Su 2015 reported that the rates of nausea and vomiting and constipation may be significantly higher in the CR oxycodone group than in the TD fentanyl group, and that there may be little to no difference in the rates of dizziness and lethargy between the groups. We judged the certainty of evidence for this outcome to be very low. We downgraded the certainty of evidence by two levels for very serious limitations to study quality due to risk of bias (arising from under‐reporting, performance bias, detection bias, attrition bias and selective reporting bias), and by one or two levels (depending on the event rates of the individual adverse events) for imprecision (arising from low or very low event rates). See Table 5.

Table 5. Controlled‐release (CR) oxycodone versus transdermal fentanyl or buprenorphine

Comparison	CR oxycodone versus TD fentanyl						CR oxycodone versus TD buprenorphine
Study	Corli 2016		Nosek 2017		Su 2015		Corli 2016		Nosek 2017
Treatment	Oxy	Fen	Oxy	Fen	Oxy	Fen	Oxy	Bup	Oxy	Bup
Severe/moderate adverse events	48.8%	50.4%			‐	‐	48.8%	60%
Breathlessness	12/129	22/127			‐	‐	12/129	30/130
Breathlessness, severe	3/129	5/127			‐	‐	3/129	6/130
Confusion	55/129	46/127			‐	‐	55/129	61/130
Confusion, severe	12/129	8/127			‐	‐	12/129	12/130
Constipation	75/129	77/127	1.91 (1.15)^a	2.77 (1.45)^a	13/42	3/38	75/129	87/130	1.91 (1.15)^a	1.76 (2.06)^a
Constipation, severe	40/129	36/127			‐	‐	40/129	39/130
Dizziness	‐	‐			4/42	3/38	‐	‐
Drowsiness	74/129	70/127			‐	‐	74/129	81/130
Drowsiness, severe	34/129	26/127			‐	‐	34/129	40/130
Dry mouth	66/129	67/127			‐	‐	66/129	73/130
Dry mouth, severe	27/129	29/127			‐	‐	27/129	30/130
Dysuria	17/129	13/127			2/42	3/38	17/129	16/130
Dysuria, severe	4/129	4/127			‐	‐	4/129	4/130
Gastralgia	21/129	26/127			‐	‐	21/129	21/130
Gastralgia, severe	6/129	4/127			‐	‐	6/129	1/130
Hallucinations	8/129	3/127			‐	‐	8/129	8/130
Hallucinations, severe	1/129	0/127			‐	‐	1/129	2/130
Itching	16/129	14/127			‐	‐	16/129	21/130
Itching, severe	2/129	3/127			‐	‐	2/129	1/130
Lethargy	‐	‐			3/42	2/38	‐	‐
Muscle spasm/myoclonus	23/129	15/127			‐	‐	23/129	24/130
Muscle spasm/myoclonus, severe	0/129	3/127			‐	‐	0/129	1/130
Nausea and vomiting	‐	‐			11/42	2/38	‐	‐
Nausea	63/129	57/127			‐	‐	63/129	59/130
Nausea, severe	22/129	16/127			‐	‐	22/129	18/130
Vomiting	29/129	29/127			‐	‐	29/129	30/130
Vomiting, severe	12/129	10/127			‐	‐	12/129	5/130
Discontinuation due to adverse events			0/16	1/15					0/16	1/17

^aBowel function index mean (standard deviation) (range 0‐10, with constipation indicated if score = 3 or above) at 28 days

‐: not reported
Bup: buprenorphine
CR: controlled‐release
Fen: fentanyl
Oxy: oxycodone
TD: transdermal

Quality of life

The evidence was very uncertain about the effect of CR oxycodone versus TD fentanyl on quality of life. For Nosek 2017, the situation was the same for this comparison as for the comparison between CR oxycodone and CR morphine reported above, with the exception that the Karnofsky Performance Scale mean score was 58 and the SD was 13.98 for the TD fentanyl group.

Su 2015 found that there may be little to no difference in quality of life as measured by the Karnofsky Performance Status (from 0 to 100) (CR oxycodone mean 75.79; TD fentanyl mean 74.05). We judged the certainty of evidence for this outcome to be very low. We downgraded the certainty of evidence by two levels for very serious limitations to study quality due to risk of bias (arising from selection, performance, detection, attrition and selective reporting bias), and by two levels for imprecision (arising from very low participant numbers).

Participant preference

None of the studies reported participant preference.

Controlled‐release oxycodone versus transdermal buprenorphine

Two studies compared CR oxycodone to TD buprenorphine (Corli 2016; Nosek 2017).

Pain intensity and pain relief

The evidence was very uncertain about the effect of CR oxycodone versus TD buprenorphine on pain intensity and pain relief. Corli 2016 compared CR oxycodone with TD buprenorphine in a four‐arm trial of 28 days' duration, which also included a TD fentanyl and a CR morphine group, and found that there may be little to no difference between CR oxycodone and TD buprenorphine groups in terms of pain intensity (Analysis 1.1; Figure 4) or pain relief (Analysis 1.2; Figure 5). Corli 2016 found that there may be little to no difference between the groups in terms of requirement for additional opioids (CR oxycodone 33/125 participants; TD buprenorphine 48/127 participants), opioid escalation index > 5% (CR oxycodone 24/125 participants; TD buprenorphine 18/127 participants), premature discontinuations for pain‐related reasons (CR oxycodone 19/125 participants; TD buprenorphine 26/127 participants), proportion of participants requiring adjuvant drugs (CR oxycodone 102/125 participants; TD buprenorphine 100/127 participants), or proportion of participants requiring switches (CR oxycodone 15/125 participants; TD buprenorphine 21/127 participants). For Nosek 2017, the situation was the same for this comparison as for the comparison between CR oxycodone and CR morphine reported above. We judged the certainty of evidence for this outcome to be very low. We downgraded the certainty of evidence by two levels for very serious limitations to study quality due to risk of bias (arising from performance bias, detection bias, attrition bias and selective reporting bias), and by one level for imprecision (arising from low participant numbers).

Adverse events

The evidence was very uncertain about the effect of CR oxycodone versus TD buprenorphine on adverse events. Corli 2016 found that there may be little to no difference between the two treatment groups in the incidence of the following adverse events, whether they were 'any degree' or 'severe': drowsiness, confusion, nausea, vomiting, constipation, dry mouth, hallucinations, gastralgia, dysuria, muscle spasm myoclonus, and itching. 'Any degree,' but not severe, breathlessness may occur more often in the TD buprenorphine group than in the CR oxycodone group. For Nosek 2017, the situation was the same for this comparison as for the comparison between CR oxycodone and CR morphine reported above. We judged the certainty of evidence for this outcome to be very low. We downgraded the certainty of evidence by two levels for very serious limitations to study quality due to risk of bias (arising from performance bias, detection bias, attrition bias and selective reporting bias), and by one or two levels (depending on the event rates for the individual adverse events) for imprecision (arising from low or very low event rates). See Table 5.

Quality of life

The evidence was very uncertain about the effect of CR oxycodone versus TD buprenorphine on quality of life. Corli 2016 did not report quality of life. For Nosek 2017, the situation was the same for this comparison as for the comparison between CR oxycodone and CR morphine reported above with the exception that the Karnofsky Performance Scale mean score was 64.67 and the SD was 10.6 for the TD buprenorphine group. We judged the certainty of evidence for this outcome to be very low. We downgraded the certainty of evidence by two levels for very serious limitations to study quality due to risk of bias (arising from performance bias, detection bias, attrition bias and selective reporting bias), and by two levels for imprecision (arising from very low participant numbers).

Participant preference

Neither study reported participant preference.

Controlled‐release oxycodone versus oral ibuprofen

One study compared CR oxycodone to oral ibuprofen (Liu 2021).

Pain intensity and pain relief

The evidence was very uncertain about the effect of CR oxycodone versus oral ibuprofen on pain relief. Liu 2021, in a parallel‐group trial of seven days' duration, found that the rates of achieving complete or significant pain relief may be higher after treatment with CR oxycodone than with oral ibuprofen (Analysis 1.2; Figure 5). One of the 32 participants in the ibuprofen group achieved no pain relief whereas this was the case for none of the 34 participants in the oxycodone group. The remaining participants in both groups achieved partial pain relief. We judged the certainty of evidence for this outcome to be very low. We downgraded the certainty of evidence by two levels for very serious limitations to study quality due to risk of bias (arising from uncertainty about allocation concealment and performance and detection bias), and by two levels for imprecision (arising from low participant/event numbers).

Adverse events

The evidence was very uncertain about the effect of CR oxycodone versus oral ibuprofen on adverse events. Inspection of Table 4 suggested that there may be little to no difference in adverse event rates between the treatment groups, which was also confirmed by the analyses reported by Liu 2021. All of the adverse events were grade I or II. No grade III or above adverse events occurred. We judged the certainty of evidence for this outcome to be very low, We downgraded the certainty of evidence by two levels for very serious limitations to study quality due to risk of bias (arising from uncertainty about allocation concealment and performance and detection bias), and by two levels for imprecision (arising from low participant/event numbers).

Quality of life

The evidence was very uncertain about the effect of CR oxycodone versus oral ibuprofen on quality of life. Liu 2021 assessed quality of life using the European Organization for Research and Treatment of Cancer Quality of Life Questionnaire‐Core15_Palliative (EORTC QLQ‐C15‐PAL) scale and the Edmonton Symptom Assessment System (ESAS) and found that there may be little to no difference in quality of life between the treatment groups (mean (SD) EORTC QLQ‐C15‐PAL scores after treatment = 20 (2.67) in the oxycodone group and 21.09 (2.37) in the ibuprofen group; mean (SD) ESAS scores after treatment = 18.82 (5.59) in the oxycodone group and 20.19 (4.3) in the ibuprofen group). We judged the certainty of evidence for this outcome to be very low. We downgraded the certainty of evidence by two levels for very serious limitations to study quality due to risk of bias (arising from uncertainty about allocation concealment and performance and detection bias), and by two levels for imprecision (arising from low participant/event numbers).

Participant preference

The study did not report participant preference.

Immediate‐release oxycodone versus immediate‐release hydromorphone

One study compared IR oxycodone to IR hydromorphone (Inoue 2018).

Pain intensity and pain relief

The evidence was very uncertain about the effect of IR oxycodone versus IR hydromorphone on pain intensity. Inoue 2018, in a parallel‐group trial of five days' duration, found that there may be little to no difference in pain intensity between the study groups (Analysis 1.1; Figure 4). We judged the certainty of evidence for this outcome to be very low. We downgraded the certainty of evidence by one level for serious limitations to study quality due to risk of bias (arising from under‐reporting of performance and detection bias), and by two levels for imprecision (arising from low participant numbers).

Adverse events

The evidence was very uncertain about the effect of IR oxycodone versus IR hydromorphone on adverse events. Inspection of Table 4 suggested that there may be little to no difference in adverse event rates between the treatment groups, which was also confirmed by the analyses reported by Inoue 2018. We judged the certainty of evidence for this outcome to be very low. We downgraded the certainty of evidence by one level for serious limitations to study quality due to risk of bias (arising from under‐reporting of performance and detection bias), and by two levels for imprecision (arising from low participant/event numbers).

Quality of life

The study did not report quality of life.

Participant preference

The study did not report participant preference.

Intravenous oxycodone versus rectal oxycodone

One study compared IV oxycodone to rectal oxycodone (Leow 1995).

Pain intensity and pain relief

The evidence was very uncertain about the effect of IV oxycodone versus rectal oxycodone on pain relief. Leow 1995 conducted a single‐dose cross‐over study in 12 participants, with each phase lasting 24 hours, and found that while IV oxycodone may be associated with faster onset of pain relief relative to rectal oxycodone, rectal oxycodone may be associated with a longer duration of pain relief compared to IV oxycodone. We judged the certainty of evidence for this outcome to be very low. We downgraded the certainty of evidence by two levels for very serious limitations to study quality due to risk of bias (arising from under‐reporting and performance bias and detection bias), and by two levels for imprecision (arising from very low participant numbers).

Adverse events

The evidence was very uncertain about the effect of IV oxycodone versus rectal oxycodone on adverse events. Leow 1995 found that there may be little to no difference in the adverse event profiles for the two study arms (see Table 6). We judged the certainty of evidence for this outcome to be very low. We downgraded the certainty of evidence by two levels for very serious limitations to study quality due to risk of bias (arising from under‐reporting, performance bias and detection bias), and by two levels for imprecision (arising from very low participant numbers).

Table 6. Intravenous (IV) oxycodone comparisons: adverse events

Comparison	IV oxycodone versus IV morphine		IV oxycodone versus rectal oxycodone		IV oxycodone followed by IR oxycodone versus IV morphine followed by IR morphine
Study	Lee 2017		Leow 1995 ^a		Kalso 1990 ^b
Treatment	Oxy	Mor	IV	Rectal	IV oxy	IR oxy	IV mor	IR mor
Any adverse events	29/34	26/32	‐	‐	‐	‐	‐	‐
Total adverse events	‐	‐	82	94	‐	‐	‐	‐
Severe adverse events	‐	‐	‐	‐	‐	‐	‐	‐
Serious adverse events	3/34	2/32	‐	‐	‐	‐	‐	‐
Unexpected adverse events	9/34	16/32
Adverse drug reaction	14/34	11/32
Abdominal discomfort	‐	‐	‐	‐	‐	‐	‐	‐
Abdominal distension	‐	‐	‐	‐	‐	‐	‐	‐
Anaemia	‐	‐	‐	‐	‐	‐	‐	‐
Anorexia, appetite loss	‐	‐	‐	‐	‐	‐	‐	‐
Asthenia	‐	‐	‐	‐	‐	‐	‐	‐
Bone marrow failure	‐	‐	‐	‐	‐	‐	‐	‐
Chest discomfort	‐	‐	‐	‐	‐	‐	‐	‐
Confusion	‐	‐	‐	‐	0/19	1/19	0/19	1/19
Constipation	13/34	6/32	‐	‐	6/19	6/19	8/19	8/19
Delirium	‐	‐	‐	‐	‐	‐	‐	‐
Diarrhoea	‐	‐	‐	‐	‐	‐	‐	‐
Dizziness or lightheadedness	‐	‐	0.54 (0.74)	0.71 (0.9)	‐	‐	‐	‐
Drowsiness, somnolence	‐	‐	0.68 (0.81)	0.79 (0.93)	7/19	4/19	4/19	5/19
Dyspnoea (serious)	0/34	1/32
Fever (serious)	1/34	0/32
Gastrointestinal disorders	22/34	16/32
General disorders and administration site	6/34	6/32
Hallucinations	‐	‐	‐	‐	0/19	0/19	2/19	3/19
Hypoproteinaemia	‐	‐	‐	‐	‐	‐	‐	‐
Insomnia	‐	‐	‐	‐	‐	‐	‐	‐
Nausea	10/34	8/32	0.02 (0.15)	0.12 (0.45)	7/19	7/19	7/19	12/19
Neutrophil count decreased/neutropenia	1/34	1/32	‐	‐	‐	‐	‐	‐
Nervous system disorders	7/34	5/32
Oedema, peripheral	‐	‐	‐	‐	‐	‐	‐	‐
Platelet count decreased	1/34	0/32	‐	‐	‐	‐	‐	‐
Pruritus	‐	‐	0.05 (0.21)	0.05 (0.21)	3/19	1/19	3/19	2/19
Pyrexia	‐	‐	‐	‐	‐	‐	‐	‐
Rash	‐	‐	‐	‐	‐	‐	‐	‐
Sedation	‐	‐	‐	‐	12/19	13/19	12/19	14/19
Skin and subcutaneous tissue disorders	5/34	4/32
Sweating or hyperhidrosis	‐	‐	0.04 (0.19)	0.07 (0.3)	4/19	2/19	1/19	1/19
Urinary retention	‐	‐	‐	‐	1/19	1/19	2/19	0/19
Urinary tract infection	‐	‐	‐	‐	‐	‐	‐	‐
Vomiting	‐	‐	0.01 (0.11)	0.01 (0.11)		‐	‐	‐
White blood cell count decreased	‐	‐	‐	‐	‐	‐	‐	‐
Other disorders	15/34	16/32
Discontinuation due to adverse events	2/34	0/32	‐	‐	‐	‐	‐	‐
Deaths due to study drug treatment	‐	‐	‐	‐	‐	‐	‐	‐

^aMean (standard deviation) ratings (out of 3) experienced during the 24 hours of drug administration, apart from the total number of adverse events which is read from the authors' Figure 3

^bThe measure is the sum of positive responses after each study period: moderate = 1, severe = 2.

‐: not reported
CR: controlled‐release
ER: extended‐release
Hyd: hydromorphone
IV: intravenous
Mor: morphine
Oxy: oxycodone
Oxymo: oxymorphone
Tap: tapentadol

Quality of life

The study did not report quality of life.

Participant preference

The study did not report participant preference.

Intravenous oxycodone versus intravenous morphine

One study compared IV oxycodone to IV morphine (Lee 2017).

Pain intensity and pain relief

The evidence was very uncertain about the effect of IV oxycodone versus IV morphine on pain intensity and pain relief. Lee 2017 compared IV oxycodone with IV morphine in a 5‐day parallel‐group trial in 66 participants, and found that there may be little to no difference in pain intensity between the two treatments (Analysis 1.1; Figure 4). There may also be little to no difference in pain relief with 78.8% and 75% of participants experiencing a reduction in pain of at least 30% from baseline to end of treatment with oxycodone and morphine, respectively, while for 63.6% and 62.5% of participants in the oxycodone and morphine groups, respectively, the reduction in pain was at least 50% (Analysis 1.2; Figure 5). We judged the certainty of evidence for this outcome to be very low. We downgraded the certainty of evidence by two levels for very serious limitations to study quality due to risk of bias (arising from under‐reporting, performance bias and detection bias), and by two levels for imprecision (arising from very low participant numbers).

Adverse events

The evidence was very uncertain about the effect of IV oxycodone versus IV morphine on adverse events. Lee 2017 found that there may be little to no difference in the adverse event profiles for the two study arms (see Table 6) although they did report "There was a difference in the unexpected AEs with more (29) events affecting 16 patients (50.0%) in the morphine group compared to 12 events in the oxycodone group affecting 9 patients (26.5%) (P value 0.049)" (page 7). We judged the certainty of evidence for this outcome to be very low. We downgraded the certainty of evidence by two levels for very serious limitations to study quality due to risk of bias (arising from under‐reporting, performance bias and detection bias), and by two levels for imprecision (arising from very low participant numbers).

Quality of life

The study did not report quality of life.

Participant preference

The evidence was very uncertain about the effect of IV oxycodone versus IV morphine on participant preference. The study did not report participant preference, but it did report treatment satisfaction, both as assessed by the participants and by the investigator. Nineteen of the 32 participants in the oxycodone group and 18 of the 31 participants in the morphine group rated their overall analgesic treatment satisfaction regarding pain as "very much improved" or much "improved", with 12 participants in each group indicating minimal improvement and one in each group indicating "no change". When this outcome was assessed by the investigator, the corresponding numbers were 26 and 24 in the oxycodone and morphine groups, respectively, rated as "very much improved" or "much improved" and five and six participants in the oxycodone and morphine groups, respectively, experiencing "minimal improvement", and again one participant in each group experiencing "no change". We downgraded the certainty of evidence by two levels for very serious limitations to study quality due to risk of bias (arising from under‐reporting, performance bias and detection bias), and by two levels for imprecision (arising from very low participant numbers).

Intravenous oxycodone followed by immediate‐release oxycodone versus intravenous morphine followed by immediate‐release morphine

One study compared IV oxycodone followed by IR oxycodone to IV morphine followed by IR morphine (Kalso 1990).

Pain intensity and pain relief

There evidence was very uncertain about the effect of IV oxycodone followed by IR oxycodone versus IV morphine followed by IR morphine on pain intensity and pain relief. In a cross‐over study comparing IV oxycodone titration (two days) followed by IR oxycodone titration (two days) with IV morphine titration (two days) followed by IR morphine titration (two days) in 19 analysed participants, Kalso 1990 found that the participants achieved equal analgesia with both drugs, but around 30% more IV oxycodone was needed compared to IV morphine and around 25% less IR oxycodone was needed than IR morphine to achieve this. We judged the certainty of evidence for this outcome to be very low. We downgraded the certainty of evidence by two levels for very serious limitations to study quality due to risk of bias (arising from under‐reporting of the domains of selection, performance, and detection bias), and by two levels for imprecision (arising from very low participant numbers).

Adverse events

The evidence was very uncertain about the effect of IV oxycodone followed by IR oxycodone versus IV morphine followed by IR morphine on adverse events. Kalso 1990 found that nausea may be more common with oral morphine treatment compared to the other three treatment modalities (see also Table 6). We judged the certainty of evidence for this outcome to be very low. We downgraded the certainty of evidence by two levels for very serious limitations to study quality due to risk of bias (arising from under‐reporting of the domains of selection, performance, and detection bias), and by two levels for imprecision (arising from very low participant numbers).

Quality of life

The study did not report quality of life.

Participant preference

The evidence was very uncertain about the effect of IV oxycodone followed by IR oxycodone versus IV morphine followed by IR morphine on participant preference. Kalso 1990 reported that 10 participants expressed no treatment preference while five participants preferred oxycodone while another five participants preferred treatment with morphine. We judged the certainty of evidence for this outcome to be very low. We downgraded the certainty of evidence by two levels for very serious limitations to study quality due to risk of bias (arising from under‐reporting of the domains of selection, performance, and detection bias), and by two levels for imprecision (arising from very low participant numbers).

Intramuscular oxycodone versus oral oxycodone

One study compared IM oxycodone to oral oxycodone (Beaver 1978a).

Pain intensity and pain relief

The evidence was very uncertain about the effect of IM oxycodone versus oral oxycodone on pain intensity and pain relief. In a single‐dose, cross‐over study, Beaver 1978a compared 5 mg and 15 mg IM oxycodone to 10 mg and 30 mg oral oxycodone in 17 participants, of whom 13 completed at least one cross‐over round of the study medications. Beaver 1978a reported that oral oxycodone may be 0.57 (95% CI 0.22 to 1.84) times as potent as IM oxycodone for pain relief and 0.78 (95% CI 0.3 to 8.82) times as potent for change in pain intensity. We judged the certainty of evidence for this outcome to be very low. We downgraded the certainty of evidence by two levels for very serious limitations to study quality due to risk of bias (arising from under‐reporting of the domains of selection and attrition bias), and by two levels for imprecision (arising from very low participant numbers).

Adverse events

The evidence was very uncertain about the effect of IM oxycodone versus oral oxycodone on adverse events. Beaver 1978a reported that the adverse effects for both oral and IM oxycodone, although infrequent, were related to dose, but otherwise provided no further details on the observed adverse effects. We judged the certainty of evidence for this outcome to be very low. We downgraded the certainty of evidence by two levels for very serious limitations to study quality due to risk of bias (arising from under‐reporting of the domains of selection and attrition bias), and by two levels for imprecision (arising from very low participant numbers).

Quality of life

The study did not report quality of life.

Participant preference

The study did not report participant preference.

Intramuscular oxycodone versus intramuscular morphine versus intramuscular codeine

Two studies (reported in one publication) compared IM oxycodone, IM morphine and IM codeine (Beaver 1978b).

Pain intensity and pain relief

The evidence was very uncertain about the effect of IM oxycodone versus IM morphine versus IM codeine on pain intensity and pain relief. In another single‐dose, cross‐over study Beaver 1978b compared 7.5 mg, 15 mg and 30 mg IM oxycodone to 8 mg, 16 mg and 32 mg IM morphine in 34 participants, of whom 28 completed at least one round of the study drugs. In this study, IM oxycodone was found to be 0.74 (95% CI 0.36 to 1.2) times as potent as IM morphine for pain relief and 0.68 (95% CI 0.32 to 1.07) times as potent as IM morphine for change in pain intensity. In a further study of similar design, Beaver 1978b compared 7.5 mg, 15 mg and 30 mg IM oxycodone to 90 mg and 180 mg IM codeine and to 16 mg IM morphine in 30 participants, of whom 26 completed at least one cross‐over round of the study medications. Beaver 1978b reported that IM oxycodone may be 10.72 (95% CI not reported) times as potent as IM codeine for pain relief and 8.44 (95% CI 2.13 to 44.69) times as potent as IM codeine for change in pain intensity. We judged the certainty of evidence for this outcome to be very low. We downgraded the certainty of evidence by two levels for very serious limitations to study quality due to risk of bias (arising from under‐reporting of the domains of selection and attrition bias), and by two levels for imprecision (arising from very low participant numbers).

Adverse events

The evidence was very uncertain about the effect of IM oxycodone versus IM morphine versus IM codeine on adverse events. Beaver 1978b noted that, in both studies, side effects typical of narcotic analgesics were observed, although not in sufficient numbers to allow meaningful analysis, and they reported no further details on adverse events. We judged the certainty of evidence for this outcome to be very low. We downgraded the certainty of evidence by two levels for very serious limitations to study quality due to risk of bias (arising from under‐reporting of the domains of selection and attrition bias), and by two levels for imprecision (arising from very low participant numbers).

Quality of life

The studies did not report quality of life.

Participant preference

The studies did not report participant preference.

Discussion

For the current update, we identified 19 additional studies for inclusion, which allowed us to perform further meta‐analyses, both on pain relief data and on more adverse event data. We were also able to examine three new treatment comparisons with oxycodone, and further assess the robustness of our primary pain intensity meta‐analyses by performing new sensitivity analyses. The main new issue in the context of this review that was identified in this update was the inclusion of 14 Chinese language studies and we have discussed the issues that arise in this context in detail already (see Excluded studies). Despite the addition of a further 19 studies in this update, the updated results and conclusions remain the same. This is because the only potentially new results were driven by the Chinese language studies and not corroborated by sensitivity analyses (e.g. for CR oxycodone superiority over CR morphine in terms of pain intensity and constipation).

Summary of main results

Overall, we included 42 studies which enrolled/randomised 4485 participants, with 3945 of these analysed for efficacy and 4176 for safety. The studies examined a number of different drug comparisons.

Four studies compared CR oxycodone to IR oxycodone, and pooled analysis of three of these studies showed that there may be little to no difference in pain intensity after treatment with either CR or IR oxycodone (SMD 0.1, 95% CI ‐0.06 to 0.26). This is also in line with the finding that none of the included studies reported that the pain intensity differed between the treatment groups. Pooled analyses of the adverse event data from three of the studies found that there may be little to no difference between the treatments in RRs for any of the adverse events, and the fourth study, which could not be included in the pooled analyses, reported no differences in adverse events either. Three of the four studies also found that there may be little to no difference in treatment acceptability between the comparisons. We noted that IR oxycodone was given every six hours rather than every four hours in these studies. This might have biased the efficacy data in favour of CR oxycodone; however, the adverse effect data suggest that giving IR oxycodone every four hours (more frequently) would have resulted in greater adverse effects, which would have mitigated advantages in efficacy.

Twenty‐four studies compared CR oxycodone to CR morphine and pooled analysis suggested that pain intensity may be lower (better) after treatment with CR morphine than CR oxycodone (SMD 0.14, 95% CI 0.01 to 0.27). However, this result was not corroborated by a sensitivity analysis that excluded the two cross‐over trials included in the overall analysis (SMD 0.12, 95% CI ‐0.02 to 0.26). Further pooled analysis showed that there may be little to no difference in the proportions of participants achieving complete or significant pain relief between CR oxycodone and CR morphine (RR 1.02, 95% CI 0.95 to 1.10). No data were available for quality of life. The evidence is very uncertain about the treatment effects on treatment acceptability and participant preference. Pooled analyses showed that, for most of the adverse events, the CIs were wide, including no effect as well as potential benefit and harm, for the comparison between CR oxycodone and CR morphine. Participants treated with CR morphine may be at about double the risk of experiencing hallucinations and insomnia and lethargy compared to participants treated with CR oxycodone, and participants treated with CR oxycodone may be at 25% lower risk of experiencing constipation. However, the latter result did not remain significant in sensitivity analyses, excluding the Chinese language studies. These findings also contrast somewhat with those reported in Lauretti 2003, which was different in design to the other studies and examined IR morphine consumption during treatment with CR oxycodone and CR morphine while keeping the ratio of CR oxycodone and CR morphine constant. Lauretti 2003 found that the participants consumed 38% more IR morphine during treatment with CR morphine than with CR oxycodone, and that CR and IR morphine may be associated with more nausea and vomiting and a similar acceptance to the study drugs compared to the combination of CR oxycodone and IR morphine, and therefore concluded that CR oxycodone combined with IR morphine is associated with superior analgesia and lower or similar rates of adverse events than a combination of CR and IR morphine.

Three studies compared CR oxycodone to TD fentanyl and pooled analysis of two of them found that there may be little to no difference in pain intensity (SMD 0.02, 95% CI ‐0.19 to 0.24) or complete/significant pain relief rates (RR 0.98, 95% CI 0.85 to 1.14) after treatment with CR oxycodone and TD fentanyl. One of the studies also found that quality of life may not differ between the treatments, but there was some disagreement between the study results in terms of adverse events with one of the studies finding that the rates of nausea and vomiting, and constipation may be higher in the CR oxycodone group than in the TD fentanyl group, whereas the other study reported that there may be little to no difference in these (and other) adverse event rates between the treatment groups.

Two studies compared CR oxycodone to TD buprenorphine but one of the studies did not report pain data in a manner where it could be included and assessed. The study that did contribute data suggested that there may be little to no difference between the treatments in terms of pain intensity, complete/significant pain relief rates and adverse event rates with the exception of 'any degree', but not severe, breathlessness, which may occur more often in the TD buprenorphine group than in the CR oxycodone group

Two studies compared CR oxycodone to ER hydromorphone. Pooled analysis showed that there may be little to no difference between the treatment groups in pain intensity (SMD 0.04, 95% CI ‐0.21 to 0.28) or in adverse event rates. The data for nausea and vomiting could not be pooled due to excessive between‐study heterogeneity and, in line with this, for both nausea and vomiting, Inoue 2017 found that the rates of nausea and vomiting may be lower in the CR oxycodone group compared to the ER hydromorphone group, whereas Yu 2014 found that there may be little to no difference in them between the interventions.

The remaining studies all compared either oxycodone in different formulations or oxycodone to different alternative opioids and none of them found any clear superiority or inferiority of oxycodone for cancer pain, neither as an analgesic agent nor in terms of adverse event rates or treatment acceptability.

Overall completeness and applicability of evidence

Although the findings of this review are applicable to the population and comparisons defined for this review, that is, adults with cancer who need treatment with strong opioids for cancer pain, they should be taken in the context that this review found 42 studies that were eligible for inclusion and these studies reported on 16 different comparisons involving oxycodone and included only 4485 participants. Moreover, for some of the outcomes (participant preference and quality of life) there were extremely few data available. To somewhat mitigate this shortfall, we reported treatment acceptability as a proxy. However, that does not change the fact that the evidence base for the effectiveness and tolerability of oxycodone (relative or absolute) for pain in adults with cancer was very limited and it did not allow us to examine the effectiveness and tolerability of oxycodone in detail through participant or treatment subgroup analyses. The current evidence base would therefore benefit from more well‐designed, large RCTs.

Quality of the evidence

The certainty of the evidence for all the outcomes was low or very low, meaning we have little confidence in the effect estimate and the true effect may be substantially different from the estimate of the effect. This is due to imprecision (low participant numbers) in some cases, and serious or very serious study limitations in all cases. In general, the assessment of the quality of the included studies was limited by a great extent of under‐reporting in the studies, especially for Chinese language studies in general. For the other studies, the participant selection items (random sequence generation and allocation concealment) were also severely under‐reported, while blinding appeared to be reasonably well undertaken overall, both in terms of treatment performance and outcome assessment. However, as is not unusual for pain research, the results were substantially compromised by attrition, with data missing from 12% of the enrolled/randomised participants for efficacy, and from just under 7% for safety. These are substantial proportions and, while it did not appear to be selective attrition, the results must be interpreted with caution, especially as they are likely to be under‐estimates given the amount of under‐reporting associated with the Chinese language studies in particular, which meant we could not be sure we had captured study dropouts/the extent of missing data fully.

Potential biases in the review process

We undertook the review according to the methods specified in our protocol, which were all in line with the recommendations of Cochrane as outlined in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011), and included a thorough search strategy designed to maximise the chances of identifying all relevant studies. Contacting authors resulted in no additional studies being identified, that is, the review therefore only contains data from published studies, some of which have not reported all the outcome data despite having apparently collected these data. The review may therefore be at some risk of publication bias, although publication bias is usually associated with positive results, and the majority of the included studies did not find differences between their treatment groups in terms of efficacy and safety. Although our meta‐analyses of pain intensity included data from cross‐over studies that were analysed as if they were parallel‐group studies, which as outlined in Unit of analysis issues results in unit of analysis error (although, in turn, this leads to an under‐weighting rather than over‐weighting of the data), the results were also subjected to sensitivity analyses that excluded the cross‐over trials, and that needed to confirm the results in order for them to be accepted as true results. In this section, we return again to the issue of whether or not we should have included Chinese databases in our search strategy to increase the applicability of our results and decrease the geographical bias associated with mainly including studies conducted in the West and published in English. As outlined already in the Excluded studies section, we would have liked to include studies from searches of Chinese databases if we could have been sufficiently confident in the integrity of the data and the correct use of the term 'random' in the context of patient allocation and if the study methods and data were reported in full. Unfortunately, this is not currently the case, but we will keep this option open in future updates of this review.

Agreements and disagreements with other studies or reviews

King 2011 conducted a systematic review without meta‐analysis that also included observational studies and concluded that, "There is no evidence from the included trials of a significant difference in analgesia or adverse effects between oxycodone and morphine or hydromorphone" (page 454). Caraceni 2011 reached a similar conclusion in their systematic review without meta‐analysis. Bekkering 2011 and Reid 2006 both included meta‐analyses in their systematic reviews and they also concluded that the effectiveness of oxycodone and morphine did not differ, although the inclusion criteria employed by Bekkering 2011 differed from ours, with Bekkering 2011 excluding cross‐over trials and including trials of chronic non‐malignant pain, whereas the publication of Reid 2006 before the trial of Mercadante 2010 precluded its inclusion. That said, the conclusions of all these reviews are all in agreement with those that we have reached in this review dealing with the same comparisons as the aforementioned reviews. In a more recent meta‐analysis that included seven trials that were also included in this review and purported to compare the analgesic effect of oxycodone and morphine on patients with moderate and advanced cancer pain, Guo 2018 concluded that "The results of this meta‐analysis demonstrate clinical non‐inferiority of morphine compared with oxycodone in alleviating cancer pain, with respect to [having] achieved a comparable clinical response whether morphine or oxycodone was used as first‐line opioid in the treatment of cancer‐related pain" (page 5). While this conclusion is also in general agreement with our review, it cannot actually be made based on the data Guo 2018 has analysed and presented because the seven included trials, which were all analysed together and treated as if they examined the same drug comparison (i.e. oxycodone versus morphine), were in fact a mix of comparisons with some of them comparing CR oxycodone with CR morphine (Corli 2016; Mercadante 2010; Riley 2015; Zecca 2016) while others compared CR oxycodone to ER hydromorphone (Inoue 2017; Yu 2014) or to ER oxymorphone (Gabrail 2004). It is unclear to us why these seven trials have been analysed together and the results presented by Guo 2018 are therefore not comparable to any of the results in this review.

Figure 1

Study flow diagram

Figure 2

Risk of bias summary: review authors' judgements about each risk of bias item for each included study

Figure 3

Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies

Figure 4

Forest plot of comparison: 1 Pain, outcome: 1.1 Pain intensity and pain relief (continuous)

Figure 5

Forest plot of comparison: 1 Pain, outcome: 1.2 Complete and/or significant pain relief (categorical)

Figure 6

Forest plot of comparison: 2 Adverse events, outcome: 2.7 Constipation

Figure 7

Forest plot of comparison: 2 Adverse events, outcome: 2.10 Drowsiness/somnolence

Figure 8

Forest plot of comparison: 2 Adverse events, outcome: 2.18 Nausea

Figure 9

Forest plot of comparison: 2 Adverse events, outcome: 2.23 Vomiting

Figure 10

Forest plot of comparison: 2 Adverse events, outcome: 2.14 Hallucinations

Analysis 1.1

Comparison 1: Pain, Outcome 1: Pain intensity and pain relief (continuous)

Analysis 1.2

Comparison 1: Pain, Outcome 2: Complete and/or significant pain relief (categorical)

Analysis 2.1

Comparison 2: Adverse events, Outcome 1: Any adverse events

Analysis 2.2

Comparison 2: Adverse events, Outcome 2: Severe adverse events

Analysis 2.3

Comparison 2: Adverse events, Outcome 3: Serious adverse events, including death

Analysis 2.4

Comparison 2: Adverse events, Outcome 4: Appetite loss/anorexia

Analysis 2.5

Comparison 2: Adverse events, Outcome 5: Asthenia

Analysis 2.6

Comparison 2: Adverse events, Outcome 6: Confusion

Analysis 2.7

Comparison 2: Adverse events, Outcome 7: Constipation

Analysis 2.8

Comparison 2: Adverse events, Outcome 8: Diarrhoea

Analysis 2.9

Comparison 2: Adverse events, Outcome 9: Dizziness/lightheadedness

Analysis 2.10

Comparison 2: Adverse events, Outcome 10: Drowsiness/somnolence

Analysis 2.11

Comparison 2: Adverse events, Outcome 11: Dry mouth

Analysis 2.12

Comparison 2: Adverse events, Outcome 12: Dysuria/uroschesis

Analysis 2.13

Comparison 2: Adverse events, Outcome 13: Fever

Analysis 2.14

Comparison 2: Adverse events, Outcome 14: Hallucinations

Analysis 2.15

Comparison 2: Adverse events, Outcome 15: Headache

Analysis 2.16

Comparison 2: Adverse events, Outcome 16: Insomnia

Analysis 2.17

Comparison 2: Adverse events, Outcome 17: Insomnia & lethargy

Analysis 2.18

Comparison 2: Adverse events, Outcome 18: Nausea

Analysis 2.19

Comparison 2: Adverse events, Outcome 19: Nausea & vomiting

Analysis 2.20

Comparison 2: Adverse events, Outcome 20: Nervousness

Analysis 2.21

Comparison 2: Adverse events, Outcome 21: Pruritus

Analysis 2.22

Comparison 2: Adverse events, Outcome 22: Sweating

Analysis 2.23

Comparison 2: Adverse events, Outcome 23: Vomiting

Analysis 2.24

Comparison 2: Adverse events, Outcome 24: Discontinuation due to adverse events

Summary of findings 1. CR oxycodone compared with IR oxycodone for cancer‐related pain in adults

Outcomes	*Illustrative comparative risks (95% CI)**		Relative effect (95% CI)	No of participants (studies)	Certainty of the evidence (GRADE)	Comments
CR oxycodone compared with IR oxycodone for cancer‐related pain in adults
Patient or population: adults with cancer‐related pain Settings: in‐ or outpatients Intervention: CR oxycodone Comparison: IR oxycodone
	Assumed risk	Corresponding risk
	IR oxycodone	CR oxycodone
Pain intensity (mean across treatment or at end of treatment; length of treatment varied across trials; various pain intensity scales; SMD)	The mean pain intensity in the CR oxycodone group was 0.12 standard deviations higher (0.1 lower to 0.34 higher) than in the IR oxycodone group		SMD 0.12 (‐0.1 to 0.24)	319 (3 studies)	⊕⊕⊝⊝ very low^a,b	‐
Constipation (Event rate during treatment, length of treatment varied across trials)	224 per 1000	159 per 1000 (101 to 253)	RR 0.71 (0.45 to 1.13)	317 (3 studies)	⊕⊝⊝⊝ very low^c,d	‐
Drowsiness/somnolence (Event rate during treatment, length of treatment varied across trials)	224 per 1000	230 per 1000 (154 to 344)	RR 1.03 (0.69 to 1.54)	317 (3 studies)	⊕⊝⊝⊝ very low^c,d	‐
Nausea (Event rate during treatment, length of treatment varied across trials)	242 per 1000	206 per 1000 (136 to 310)	RR 0.85 (0.56 to 1.28)	317 (3 studies)	⊕⊝⊝⊝ very low^c,d	‐
Vomiting (Event rate during treatment, length of treatment varied across trials)	174 per 1000	115 per 1000 (66 to 200)	RR 0.66 (0.38 to 1.15)	317 (3 studies)	⊕⊝⊝⊝ very low^c,d	‐
Quality of life	No data available, but there appeared to be no difference in treatment acceptability between the treatment groups (measured on various scales, not pooled); 578 participants (3 studies); quality of the evidence low.^a
Participant preference	No data available		‐	‐	‐	‐
The assumed risk* is reported as the observed risk in the control group across studies. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; CR: controlled‐release; IR: immediate‐release; RR: risk ratio; SMD: standardised mean difference.
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate; the true effect is likely to be close to the estimate of effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited; the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate; the true effect is likely to be substantially different from the estimate of effect.
^a Downgraded twice for very serious limitations to study quality due to risk of bias (attrition bias and under‐reporting of the domain of selection bias) ^b Downgraded once for imprecision due to low event rates ^c Downgraded twice for imprecision due to very low event rates ^d Downgraded twice for very serious limitations to study quality due to risk of bias (performance/detection bias, and inadequate titration and under‐reporting of the domains of selection, performance, detection and attrition bias, and whether the participants were adequately titrated)

Summary of findings 1. CR oxycodone compared with IR oxycodone for cancer‐related pain in adults

Summary of findings 2. CR oxycodone compared with CR morphine for cancer‐related pain in adults

Outcomes	*Illustrative comparative risks (95% CI)**		Relative effect (95% CI)	No of participants (studies)	Certainty of the evidence (GRADE)	Comments
CR oxycodone compared with CR morphine for cancer‐related pain in adults
Patient or population: adults with cancer‐related pain Settings: in‐ or outpatients Intervention: CR oxycodone Comparison: CR morphine
	Assumed risk	Corresponding risk
	CR morphine	CR oxycodone
Pain intensity (various pain intensity scales; SMD)	The mean pain intensity in the CR oxycodone group was 0.14 standard deviations higher (0.01 lower to 0.27 higher) than in the CR morphine group		SMD 0.14 (0.01 to 0.27)	882 (7 studies)	⊕⊕⊝⊝ low^a	This estimate did not include the data from the Chinese language studies (n = 2) as sensitivity analyses indicated they were not comparable. Converting the SMD as a difference in Brief Pain Inventory scores (0 to 10 numerical rating scale from no pain to worst pain imaginable) between the treatments gave an estimated difference of 0.27 between the treatments, which was not clinically significant.
Constipation ‐ all available data (Event rate during treatment, length of treatment varied across trials)	322 per 1000	241 per 1000 (212 to 277)	RR 0.75 (0.66 to 0.86)	1894 (18 studies)	⊕⊕⊝⊝ low^b	This estimate did include the data from the Chinese language studies.
Constipation ‐ only English language studies (Event rate during treatment, length of treatment varied across trials)	355 per 1000	348 per 1000 (291 to 412)	RR 0.98 (0.82 to 1.16)	797 (5 studies)	⊕⊝⊝⊝ very low^b,c	This estimate did not include the data from the Chinese language studies (n = 13) as sensitivity analyses indicated they were not comparable.
Drowsiness/somnolence (Event rate during treatment, length of treatment varied across trials)	228 per 1000	201 per 1000 (169 to 239)	RR 0.88 (0.74 to 1.05)	1486 (15 studies)	⊕⊕⊝⊝ low^b	This estimate did include the data from the Chinese language studies as sensitivity analyses indicated they were comparable.
Nausea (Event rate during treatment, length of treatment varied across trials)	231 per 1000	215 per 1000 (178 to 259)	RR 0.93 (0.77 to 1.12)	1388 (13 studies)	⊕⊕⊝⊝ low^b	This estimate did include the data from the Chinese language studies as sensitivity analyses indicated they were comparable.
Vomiting (Event rate during treatment, length of treatment varied across trials)	157 per 1000	127 per 1000 (99 to 163)	RR 0.81 (0.63 to 1.04)	1388 (13 studies)	⊕⊝⊝⊝ very low^b,c	This estimate did include the data from the Chinese language studies as sensitivity analyses indicated they were comparable
Quality of life	No data available, but CR oxycodone appeared to be associated with similar or lower treatment acceptability than CR morphine (measured on various scales, not pooled); 149 participants (3 studies); quality of the evidence verylow^d,e					‐
Participant preference (end of treatment in a cross‐over trial with each phase lasting seven days)	8/23 participants preferred CR oxycodone while 11/23 participants preferred treatment with CR morphine.		‐	23 (1 study)	⊕⊝⊝⊝ very low^d,e	‐
The assumed risk* is reported as the observed risk in the control group across studies. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; CR: controlled‐release; RR: risk ratio; SMD: standardised mean difference.
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate; the true effect is likely to be close to the estimate of effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited; the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate; the true effect is likely to be substantially different from the estimate of effect.
^a Downgraded twice for very serious limitations to study quality due to risk of bias (performance/detection bias and attrition bias and under‐reporting of the domain of selection bias) ^b Downgraded twice for very serious limitations to study quality due to risk of bias (performance/detection bias and/or attrition bias and under‐reporting) ^c Downgraded once for imprecision due to low event rates/participant numbers ^d Downgraded twice for imprecision due to very low event rates/participant numbers ^e Downgraded twice for very serious limitations to study quality due to risk of bias (attrition bias and under‐reporting of the domain of selection bias)

Summary of findings 2. CR oxycodone compared with CR morphine for cancer‐related pain in adults

Table 1. Controlled‐release (CR) oxycodone versus immediate‐release (IR) oxycodone: adverse events

Comparison	CR oxycodone versus IR oxycodone
Study	Kaplan 1998		Parris 1998 ^a		Salzman 1999		Stambaugh 2001
Treatment	CR	IR	CR	IR	CR	IR	CR	IR
Any adverse events	‐	‐	38/54‐55	38/54‐55	‐	‐	10/30	10/30
Total adverse events	109	186	138	142	‐	‐	‐	‐
Abdominal pain	‐	‐	3/54‐55	1/54‐55	‐	‐	‐	‐
Anxiety	0/78	4/82	‐	‐	‐	‐	‐	‐
Asthenia	3/78	8/82	‐	‐	2/24	1/24	2/30	2/30
Confusion	‐	‐	0/54‐55	2/54‐55	3/24	2/24	‐	‐
Constipation	9/78	17/82	12/54‐55	10/54‐55	4/24	9/24	1/30	1/30
Dizziness, lightheadedness	5/78	11/82	8/54‐55	10/54‐55	2/24	0/24	3/30	3/30
Drowsiness, somnolence	14/78	17/82	13/54‐55	12/54‐55	9/24	7/24	3/30	2/30
Dry mouth	3/78	5/82	4/54‐55	3/54‐55	3/24	1/24	1/30	1/30
Headache	0/78	6/82	7/54‐55	3/54‐55	1/24	1/24	‐	‐
Insomnia	2/78	4/82	3/54‐55	1/54‐55	‐	‐	‐	‐
Nausea	14/78	21/82	11/54‐55	13/54‐55	7/24	5/24	4/30	3/30
Nervousness	3/78	5/82	‐	‐	2/24	4/24	0/30	1/30
Postural hypotension	‐	‐	‐	‐	5/24	4/24	‐	‐
Pruritus	2/78	4/82	7/54‐55	5/54‐55	4/24	0/24	1/30	2/30
Sweating	4/78	3/82	1/54‐55	5/54‐55	‐	‐	2/30	1/30
Vomiting	8/78	14/82	5/54‐55	11/54‐55	5/24	3/24	2/30	0/30
Discontinuation due to adverse events	6/78	10/82	4/54‐55	7/54‐55	1/24	2/24	‐	‐
^a Total number of participants for safety evaluation = 109. Not clear which group had 55 and 54 participants, respectively. ‐: not reported CR: controlled‐release IR: immediate‐release

Table 1. Controlled‐release (CR) oxycodone versus immediate‐release (IR) oxycodone: adverse events

Table 2. Controlled‐release (CR) oxycodone versus CR morphine (English‐language studies): adverse events

Comparison	CR oxycodone versus CR morphine
Study	Bruera 1998		Corli 2016		Heiskanen 1997		Lauretti 2003		Mercadante 2010 ^a		Mucci‐LoRusso 1998		Nosek 2017		Riley 2014		Zhang 2014			Zecca 2016
Treatment	Oxy	Mor	Oxy	Mor	Oxy	Mor	Oxy	Mor	Oxy	Mor	Oxy	Mor	Oxy	Mor	Oxy	Mor	Oxy	Mor	MS Contin	Oxy	Mor
Any adverse events	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	40/48	39/52	‐	‐	‐	‐	25/57	34/57	31/57	77/81	79/94
Severe/moderate adverse events	‐	‐	48.8%	58.9%	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐
Abnormal dreams	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	3/81	1/72	‐	‐	‐	‐	‐
Anorexia, appetite loss	‐	‐	‐	‐	0/27	1/27	14/22	13/22	‐	‐	‐	‐	‐	‐	1/81	0/72	‐	‐	‐	‐	‐
Chills	‐	‐	‐	‐	1/27	0/27	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐
Confusion ‐ serious	‐	‐	55/129 12/129	59/129 20/129	‐	‐	‐	‐	0.37 (0.49)	0.25 (0.44)	‐	‐	‐	‐	7/81 3/81	2/72 0/72	‐	‐	‐	12/85	12/88
Constipation ‐ serious	‐	‐	75/129 40/129	82/129 50/129	18/27	14/27	4/22	5/22	0.63 (0.68)	0.7 (0.92)	10/48	10/52	1.91 (1.15)^b	2.14 (1.01)^b	18/81 2/81	24/72 5/72	6/57	3/57	5/57	30/85	22/87
Decreased mobility	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	0/81	2/72	‐	‐	‐	‐	‐
Depression	‐	‐	‐	‐	1/27	0/27	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐
Diarrhoea	‐	‐	‐	‐	2/27	2/27	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐
Dizziness, lightheadedness	‐	‐	‐	‐	6/27	6/27	‐	‐	‐	‐	4/48	7/52	‐	‐	3/81	2/72	2/57	3/57	4/57	‐	‐
Double vision	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	0/81	1/72	‐	‐	‐	‐	‐
Drowsiness, somnolence ‐ serious (with/without hallucinations)	‐	‐	74/129 34/129	79/129 38/129	‐	‐	7/22	11/22	0.37 (0.6)	0.35 (0.59)	7/48	10/52	‐	‐	12/81 1/81	13/72 0/72	1/57	1/57	1/57	27/85	31/88
Drunken feeling	‐	‐	‐	‐	1/27	1/27	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐
Dry mouth ‐ severe	‐	‐	66/129 27/129	66/129 31/129	12/27	15/27	3/22	2/22	0.63 (0.68)	0.6 (0.68)	1/48	7/52	‐	‐	3/81	2/72	‐	‐	‐	19/85	14/88
Dyspnoea ‐severe	‐	‐	12/129 3/129	17/129 4/129	2/27	2/27	0/22	0/22	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐
Dysuria ‐ severe	‐	‐	17/129 4/129	22/129 2/129	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	0/57	2/57	1/57	‐	‐
Extrasystoles	‐	‐	‐	‐	1/27	0/27	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐
Faecal incontinence	‐	‐	‐	‐	1/27	1/27	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐
Fall	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	0/81	3/72	‐	‐	‐	‐	‐
Feeling abnormal	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	0/81	1/72	‐	‐	‐	‐	‐
Flatus	‐	‐	‐	‐	0/27	1/27	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐
Gastralgia ‐ severe	‐	‐	21/129 6/129	24/129 3/129	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐
Hallucinations ‐ severe	‐	‐	8/129 1/129	17/129 6/129	‐	‐	0/22	0/22	‐	‐	0/48	2/52	‐	‐	3/81	4/72	‐	‐	‐	3/91	4/94
Hollow feeling	‐	‐	‐	‐	1/27	0/27	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐
Lethargy	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	1/81	0/72	‐	‐	‐	‐	‐
Memory impairment	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	1/81	1/72	‐	‐	‐	‐	‐
Muscle twitches	‐	‐	‐	‐	1/27	1/27	‐	‐	‐	‐	‐	‐	‐	‐	0/81	2/72	‐	‐	‐	‐	‐
Muscle spasm myoclonus ‐ severe	‐	‐	23/129 0/129	14/129 6/129	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐
Nausea ‐ serious (with vomiting)/severe	12.3	13.9	63/129 22/129	64/129 19/129	14/27	16/27	1/22	8/22	0.84 (0.9)	0.6 (0.75)	6/48	8/52	‐	‐	10/81 1/81	6/72 0/72	11/57	15/57	14/57	18/85	13/88
Nightmares	‐	‐	‐	‐	0/27	3/27	‐	‐	‐	‐	‐	‐	‐	‐	2/81	0/72	‐	‐	‐	‐
Pain	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	0/81	1/72	‐	‐	‐	‐	‐
Paresthesia	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	1/81	0/72	‐	‐	‐	‐	‐
Pruritus ‐ severe	‐	‐	16/129 2/129	24/129 3/129	10/27	7/27	1/22	1/22	‐	‐	4/48	5/52	‐	‐	3/81	2/72	‐	‐	‐	6/85	5/88
Sedation	21.4	25	‐	‐	16/27	18/27	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐
Sensation of empty head	‐	‐	‐	‐	‐	‐	1/22	0/11	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐
Slow speech	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	1/81	0/72	‐	‐	‐	‐	‐
Sweating, hyperhidrosis	‐	‐	‐	‐	12/27	9/27	‐	‐	‐	‐	‐	‐	‐	‐	2/81	0/72	‐	‐	‐	‐	‐
Serious toxicity secondary to infection	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	1/81	0/72	‐	‐	‐	‐	‐
Urinary hesitation	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	0/81	1/72	‐	‐	‐	‐	‐
Visual impairment	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	1/81	0/72	‐	‐	‐	‐	‐
Vomiting ‐ severe	‐	‐	29/129 12/129	35/129 12/129	5/27	10/27	0/22	7/22	‐	‐	6/48	5/52	‐	‐	9/81	4/72	5/57	10/57	6/57	11/85	10/88
Discontinuation due to adverse events	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	3/48	6/52	0/16	1/14	‐	‐	0/57	0/57	0/57	6/91	3/94
Unexpected serious adverse events	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	2/81	7/72	‐	‐	‐	‐	‐
^a Mean (SD) ratings (out of 3) experienced during week 4 ^bBowel function index mean (standard deviation) (range 0‐10, with constipation indicated if score = 3 or above) at 28 days ‐: not reported CR: controlled‐release Mor: morphine Oxy: oxycodone

Table 2. Controlled‐release (CR) oxycodone versus CR morphine (English‐language studies): adverse events

Table 3. Controlled‐release (CR) oxycodone versus CR morphine (Chinese‐language studies): adverse events

Comparison	CR oxycodone versus CR morphine
Study	Treat‐ ment	Adverse event (AE)
Study	Treat‐ ment	Any AE	Se‐ ve‐ re AE	Ab‐ domi‐ nal dis‐ ten‐ sion	Ano‐ rexia/ appe‐ tite loss	Con‐ sti‐ pa‐ tion	Dizzi‐ ness	Dizzi‐ ness & fa‐ tigue	Drow‐ si‐ ness	Drow‐ siness & fa‐ tigue	Dry mouth	Dys‐ uria/ uro‐ sche‐ sis	Insa‐ nity/ men‐ tal disor‐ ders	In‐ som‐ nia & le‐ thar‐ gy	Nau‐ sea	Nau‐ sea & vo‐ mi‐ ting	Pru‐ ritis	Swea‐ ting	Res‐ pira‐ tory de‐ pres‐ sion	Vo‐ mi‐ ting	Dis‐ con‐ tinu‐ ation due to AE
Cao 2015	Oxy	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	3/65	‐	‐	‐	‐	1/65	‐
Cao 2015	Mor	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	11/65	‐	‐	‐	‐	7/65	‐
Gao 2012	Oxy	‐	‐	‐	‐	2/30	3/30	‐	1/30	‐	‐	‐	‐	‐	4/30	‐	‐	‐	‐	2/30	‐
Gao 2012	Mor	‐	‐	‐	‐	5/28	2/28	‐	2/28	‐	‐	‐	‐	‐	3/28	‐	‐	‐	‐	3/28	‐
Li 2013	Oxy	16/42	‐	‐	‐	9/42	‐	‐	4/42	‐	‐	0/42	‐	‐	3/42	‐	‐	‐	‐	0/42	‐
Li 2013	Mor	24/40	‐	‐	‐	12/40	‐	‐	5/40	‐	‐	1/40	‐	‐	4/40	‐	‐	‐	‐	2/40	‐
Ren 2012	Oxy	‐	0/40	‐	2/40	12/40	7/40	‐	5/40	‐	‐	‐	‐	‐	9/40	‐	2/40	‐	‐	13/40	‐
Ren 2012	Mor	‐	0/40	‐	1/40	22/40	6/40	‐	6/40	‐	‐	‐	‐	‐	8/40	‐	2/40	‐	‐	14/40	‐
Song 2015	Oxy	‐	‐	‐	‐	5/55	‐	‐	‐	‐	‐	‐	1/55	3/55	‐	6/55	2/55	‐	‐	‐	‐
Song 2015	Mor	‐	‐	‐	‐	13/55	‐	‐	‐	‐	‐	‐	7/55	10/55	‐	15/55	4/55	‐	‐	‐	‐
Sun 2013 ^a	Oxy	64/102	‐	‐	‐	19/102	‐	‐	‐	‐	1/102	0/102	2/102	10/102	‐	25/102	3/102	‐	‐	‐	‐
Sun 2013 ^a	Mor	99/102	‐	‐	‐	36/102	‐	‐	‐	‐	4/102	3/102	6/102	17/102	‐	21/102	5/102	‐	‐	‐	‐
Tu 2015	Oxy	‐	‐	‐	‐	5/43	3/43	‐	‐	2/43	‐	1/43	‐	‐	‐	5/43	‐	‐	‐	‐	‐
Tu 2015	Mor	‐	‐	‐	‐	12/43	3/43	‐	‐	3/43	‐	1/43	‐	‐	‐	4/43	‐	‐	‐	‐	‐
Wang 2008	Oxy	‐	‐	‐	‐	9/30	‐	1/30	2/30	‐	‐	1/30	‐	‐	4/30	‐	‐	‐	0/30	3/30	0/30
Wang 2008	Mor	‐	‐	‐	‐	18/30	‐	2/30	4/30	‐	‐	2/30	‐	‐	5/30	‐	‐	‐	0/30	4/30	0/30
Xie 2018	Oxy	3/48	‐	0/48	‐	1/48	1/48	‐	‐	‐	‐	‐	‐	‐	‐	1/48	‐	‐	‐	‐	‐
Xie 2018	Mor	10/47	‐	2/47	‐	2/47	3/47	‐	‐	‐	‐	‐	‐	‐	‐	3/47	‐	‐	‐	‐	‐
Ye 2012	Oxy	18/42	‐	‐	‐	3/42	2/42	‐	2/42	‐	‐	1/42	‐	‐	7/42	‐	‐	‐	‐	3/42	‐
Ye 2012	Mor	28/41	‐	‐	‐	5/41	2/41	‐	3/41	‐	‐	1/41	‐	‐	10/41	‐	‐	‐	‐	7/41	‐
Yu 2007	Oxy	‐	‐	‐	2/15	5/15	5/15	‐	2/15	‐	‐	‐	‐	‐	1/15	‐	2/15	‐	‐	2/15	‐
Yu 2007	Mor	‐	‐	‐	3/15	7/15	4/15	‐	2/15	‐	‐	‐	‐	‐	2/15	‐	3/15	‐	‐	3/15	‐
Yu 2009	Oxy	‐	‐	‐	‐	7/32	8/32	‐	5/32	‐	‐	‐	‐	‐	‐	10/32	‐	‐	‐	‐	1/32
Yu 2009	Mor	‐	‐	‐	‐	15/30	7/30	‐	4/30	‐	‐	‐	‐	‐	‐	18/30	‐	‐	‐	‐	3/30
Zhang 2011	Oxy	‐	‐	‐	‐	15/35	‐	‐	3/35	‐	‐	‐	‐	‐	10/35	‐	‐	2/35	‐	4/35	0/35
Zhang 2011	Mor	‐	‐	‐	‐	13/32	‐	‐	1/32	‐	‐	‐	‐	‐	10/32	‐	‐	0/32	‐	4/32	0/32
Zhang 2016a	Oxy	6/40	‐	‐	‐	0/40	1/40	‐	2/40	‐	‐	‐	‐	‐	‐	3/40	‐	‐	‐	‐	‐
	Mor sulfate	14/40	‐	‐	‐	2/40	5/40	‐	4/40	‐	‐	‐	‐	‐	‐	3/40	‐	‐	‐	‐	‐
	Mor hydro‐ chloride	13/40	‐	‐	‐	1/40	4/40	‐	5/40	‐	‐	‐	‐	‐	‐	3/40	‐		‐	‐	‐
^aSun 2013 also reported that 4/102 participants in the CR oxycodone group and 8/102 participants in the CR morphine group experienced "Other adverse events". ‐: not reported AE: adverse event CR: controlled‐release Mor: morphine Oxy: oxycodone

Table 3. Controlled‐release (CR) oxycodone versus CR morphine (Chinese‐language studies): adverse events

Table 4. Other oral oxycodone comparisons: adverse events

Comparison	CR oxycodone versus CR hydromorphone		CR oxycodone versus ER hydromorphone				CR oxycodone versus ER oxymorphone		CR oxycodone versus ER tapentadol		CR oxycodone versus oral ibuprofen		IR oxycodone versus IR hydromorphone
Study	Hagen 1997		Inoue 2017		Yu 2014		Gabrail 2004		Imanaka 2013		Liu 2021		Inoue 2018
Treatment	Oxy	Hyd	Oxy	Hyd	Oxy	Hyd	Oxy	Oxymo	Oxy	Tap	Oxy	Ibu	Oxy	Hyd
Any adverse events	‐	‐	77/92	71/88	117/126	111/128	‐	‐	155/172	147/168	‐	‐	65/84	73/88
Total adverse events	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐
Severe adverse events	‐	‐	10/92	7/88	30/126	30/128	‐	‐	‐	‐	0/34	0/32	‐	‐
Serious adverse events (including death)	‐	‐	14/92	11/88	18/126	11/128	‐	‐	‐	‐	0/34	0/32	8/84	7/88
Serious adverse events (including death), considered related to study drugs	‐	‐	6/92	4/88	‐	‐	‐	‐	‐	‐	0/34	0/32	3/84	4/88
Abdominal discomfort	‐	‐	‐	‐	7/126	4/128	‐	‐	‐	‐	5/34	10/32	‐	‐
Abdominal distension	‐	‐	‐	‐	7/126	7/128	‐	‐	‐	‐	‐	‐	‐	‐
Anaemia	‐	‐	‐	‐	14/126	14/128	‐	‐	‐	‐	‐	‐	‐	‐
Anorexia, appetite loss	‐	‐	3/92	7/88	21/126	20/128	‐	‐	24/172	23/168	‐	‐	‐	‐
Asthenia	‐	‐	‐	‐	9/126	11/128	‐	‐	‐	‐	‐	‐	‐	‐
Bone marrow failure	‐	‐	‐	‐	9/126	9/128	‐	‐	‐	‐	‐	‐	‐	‐
Chest discomfort	‐	‐	‐	‐	6/126	9/128	‐	‐	‐	‐	‐	‐	‐	‐
Confusion	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐
Constipation	‐	‐	14/92	11/88	45/126	43/128	19/41	21/43	64/172	51/168	9/34	3/32	19/84	21/88
Delirium	‐	‐	‐	‐	‐	‐	‐	‐	6/172	10/168	‐	‐	5/84	2/88
Diarrhoea	‐	‐	17/92	16/88	9/126	12/128	‐	‐	19/172	11/168	‐	‐	9/84	12/88
Dizziness or lightheadedness	‐	‐	5/92	6/88	22/126	21/128	9/41	7/43	‐	‐	10/34	4/32	‐	‐
Drowsiness, somnolence	28/31	19/31	18/92	23/88	‐	‐	‐	‐	36/172	29/168	3/34	0/32	21/84	23/88
Fever	‐	‐	5/92	7/88	27/126	24/128	‐	‐	‐	‐	‐	‐	‐	‐
Hallucinations	0/31	2/31	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐
Hypoproteinaemia	‐	‐	‐	‐	5/126	9/128	‐	‐	‐	‐	‐	‐	‐	‐
Insomnia	‐	‐	‐	‐	‐	‐	‐	‐	11/172	9/168	‐	‐	‐	‐
Malaise	‐	‐	6/92	3/88	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐
Nausea	15 (3)^a	13 (3)^a	21/92	36/88	45/126	43/128	15/41	17/43	61/172	48/168	3/34	0/32	14/84	14/88
Neutrophil count decreased	‐	‐	‐	‐	5/126	7/128	‐	‐	‐	‐	‐	‐	‐	‐
Oedema, peripheral	‐	‐	‐	‐	6/126	11/128	‐	‐	‐	‐	‐	‐	‐	‐
Platelet count decreased	‐	‐	‐	‐	7/126	8/128	‐	‐	‐	‐	‐	‐	‐	‐
Pruritus	‐	‐	‐	‐	‐	‐	8/41	13/43	‐	‐	‐	‐	‐	‐
Rash	‐	‐	‐	‐	4/126	7/128	‐	‐	‐	‐	‐	‐	‐	‐
Respiratory depression	‐	‐	0/92	0/88	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐
Sedation	24 (4)^a	18 (3)^a	‐	‐	‐	‐	13/41	18/43	‐	‐	‐	‐	‐	‐
Sweating or hyperhidrosis	‐	‐	‐	‐	8/126	3/128	9/41	12/43	‐	‐	‐	‐	‐	‐
Transaminases elevated	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	2/34	7/32	‐	‐
Urinary retention	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐
Urinary tract infection	‐	‐	‐	‐	7/126	4/128	‐	‐	‐	‐	‐	‐	‐	‐
Urination difficulty	‐	‐	‐	‐	‐	‐	‐	‐	‐	‐	1/34	0/32	‐	‐
Vomiting	‐	‐	16/92	32/88	47/126	43/128	7/41	5/43	41/172	42/168	‐	‐	15/84	17/88
White blood cell count decreased	‐	‐	‐	‐	17/126	13/128	‐	‐	‐	‐	‐	‐	‐	‐
Discontinuation due to adverse events	‐	‐	14/92	10/88	18/126	19/128	‐	‐	29/172	22/168	‐	‐	8/84	6/88
Deaths due to study drug treatment	‐	‐	‐	‐	0/126	0/128	‐	‐	‐	‐	0/34	0/32	‐	‐
^aMean (standard error) visual analogue scale across all days ‐: not reported CR: controlled‐release ER: extended‐release Hyd: hydromorphone Ibu: Ibuprofen IR: Immediate‐release IV: intravenous Mor: morphine Oxy: oxycodone Oxymo: oxymorphone Tap: tapentadol

Table 4. Other oral oxycodone comparisons: adverse events

Table 5. Controlled‐release (CR) oxycodone versus transdermal fentanyl or buprenorphine

Comparison	CR oxycodone versus TD fentanyl						CR oxycodone versus TD buprenorphine
Study	Corli 2016		Nosek 2017		Su 2015		Corli 2016		Nosek 2017
Treatment	Oxy	Fen	Oxy	Fen	Oxy	Fen	Oxy	Bup	Oxy	Bup
Severe/moderate adverse events	48.8%	50.4%			‐	‐	48.8%	60%
Breathlessness	12/129	22/127			‐	‐	12/129	30/130
Breathlessness, severe	3/129	5/127			‐	‐	3/129	6/130
Confusion	55/129	46/127			‐	‐	55/129	61/130
Confusion, severe	12/129	8/127			‐	‐	12/129	12/130
Constipation	75/129	77/127	1.91 (1.15)^a	2.77 (1.45)^a	13/42	3/38	75/129	87/130	1.91 (1.15)^a	1.76 (2.06)^a
Constipation, severe	40/129	36/127			‐	‐	40/129	39/130
Dizziness	‐	‐			4/42	3/38	‐	‐
Drowsiness	74/129	70/127			‐	‐	74/129	81/130
Drowsiness, severe	34/129	26/127			‐	‐	34/129	40/130
Dry mouth	66/129	67/127			‐	‐	66/129	73/130
Dry mouth, severe	27/129	29/127			‐	‐	27/129	30/130
Dysuria	17/129	13/127			2/42	3/38	17/129	16/130
Dysuria, severe	4/129	4/127			‐	‐	4/129	4/130
Gastralgia	21/129	26/127			‐	‐	21/129	21/130
Gastralgia, severe	6/129	4/127			‐	‐	6/129	1/130
Hallucinations	8/129	3/127			‐	‐	8/129	8/130
Hallucinations, severe	1/129	0/127			‐	‐	1/129	2/130
Itching	16/129	14/127			‐	‐	16/129	21/130
Itching, severe	2/129	3/127			‐	‐	2/129	1/130
Lethargy	‐	‐			3/42	2/38	‐	‐
Muscle spasm/myoclonus	23/129	15/127			‐	‐	23/129	24/130
Muscle spasm/myoclonus, severe	0/129	3/127			‐	‐	0/129	1/130
Nausea and vomiting	‐	‐			11/42	2/38	‐	‐
Nausea	63/129	57/127			‐	‐	63/129	59/130
Nausea, severe	22/129	16/127			‐	‐	22/129	18/130
Vomiting	29/129	29/127			‐	‐	29/129	30/130
Vomiting, severe	12/129	10/127			‐	‐	12/129	5/130
Discontinuation due to adverse events			0/16	1/15					0/16	1/17
^aBowel function index mean (standard deviation) (range 0‐10, with constipation indicated if score = 3 or above) at 28 days ‐: not reported Bup: buprenorphine CR: controlled‐release Fen: fentanyl Oxy: oxycodone TD: transdermal

Table 5. Controlled‐release (CR) oxycodone versus transdermal fentanyl or buprenorphine

Table 6. Intravenous (IV) oxycodone comparisons: adverse events

Comparison	IV oxycodone versus IV morphine		IV oxycodone versus rectal oxycodone		IV oxycodone followed by IR oxycodone versus IV morphine followed by IR morphine
Study	Lee 2017		Leow 1995 ^a		Kalso 1990 ^b
Treatment	Oxy	Mor	IV	Rectal	IV oxy	IR oxy	IV mor	IR mor
Any adverse events	29/34	26/32	‐	‐	‐	‐	‐	‐
Total adverse events	‐	‐	82	94	‐	‐	‐	‐
Severe adverse events	‐	‐	‐	‐	‐	‐	‐	‐
Serious adverse events	3/34	2/32	‐	‐	‐	‐	‐	‐
Unexpected adverse events	9/34	16/32
Adverse drug reaction	14/34	11/32
Abdominal discomfort	‐	‐	‐	‐	‐	‐	‐	‐
Abdominal distension	‐	‐	‐	‐	‐	‐	‐	‐
Anaemia	‐	‐	‐	‐	‐	‐	‐	‐
Anorexia, appetite loss	‐	‐	‐	‐	‐	‐	‐	‐
Asthenia	‐	‐	‐	‐	‐	‐	‐	‐
Bone marrow failure	‐	‐	‐	‐	‐	‐	‐	‐
Chest discomfort	‐	‐	‐	‐	‐	‐	‐	‐
Confusion	‐	‐	‐	‐	0/19	1/19	0/19	1/19
Constipation	13/34	6/32	‐	‐	6/19	6/19	8/19	8/19
Delirium	‐	‐	‐	‐	‐	‐	‐	‐
Diarrhoea	‐	‐	‐	‐	‐	‐	‐	‐
Dizziness or lightheadedness	‐	‐	0.54 (0.74)	0.71 (0.9)	‐	‐	‐	‐
Drowsiness, somnolence	‐	‐	0.68 (0.81)	0.79 (0.93)	7/19	4/19	4/19	5/19
Dyspnoea (serious)	0/34	1/32
Fever (serious)	1/34	0/32
Gastrointestinal disorders	22/34	16/32
General disorders and administration site	6/34	6/32
Hallucinations	‐	‐	‐	‐	0/19	0/19	2/19	3/19
Hypoproteinaemia	‐	‐	‐	‐	‐	‐	‐	‐
Insomnia	‐	‐	‐	‐	‐	‐	‐	‐
Nausea	10/34	8/32	0.02 (0.15)	0.12 (0.45)	7/19	7/19	7/19	12/19
Neutrophil count decreased/neutropenia	1/34	1/32	‐	‐	‐	‐	‐	‐
Nervous system disorders	7/34	5/32
Oedema, peripheral	‐	‐	‐	‐	‐	‐	‐	‐
Platelet count decreased	1/34	0/32	‐	‐	‐	‐	‐	‐
Pruritus	‐	‐	0.05 (0.21)	0.05 (0.21)	3/19	1/19	3/19	2/19
Pyrexia	‐	‐	‐	‐	‐	‐	‐	‐
Rash	‐	‐	‐	‐	‐	‐	‐	‐
Sedation	‐	‐	‐	‐	12/19	13/19	12/19	14/19
Skin and subcutaneous tissue disorders	5/34	4/32
Sweating or hyperhidrosis	‐	‐	0.04 (0.19)	0.07 (0.3)	4/19	2/19	1/19	1/19
Urinary retention	‐	‐	‐	‐	1/19	1/19	2/19	0/19
Urinary tract infection	‐	‐	‐	‐	‐	‐	‐	‐
Vomiting	‐	‐	0.01 (0.11)	0.01 (0.11)		‐	‐	‐
White blood cell count decreased	‐	‐	‐	‐	‐	‐	‐	‐
Other disorders	15/34	16/32
Discontinuation due to adverse events	2/34	0/32	‐	‐	‐	‐	‐	‐
Deaths due to study drug treatment	‐	‐	‐	‐	‐	‐	‐	‐
^aMean (standard deviation) ratings (out of 3) experienced during the 24 hours of drug administration, apart from the total number of adverse events which is read from the authors' Figure 3 ^bThe measure is the sum of positive responses after each study period: moderate = 1, severe = 2. ‐: not reported CR: controlled‐release ER: extended‐release Hyd: hydromorphone IV: intravenous Mor: morphine Oxy: oxycodone Oxymo: oxymorphone Tap: tapentadol

Table 6. Intravenous (IV) oxycodone comparisons: adverse events

Comparison 1. Pain

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1.1 Pain intensity and pain relief (continuous) Show forest plot	23		Std. Mean Difference (IV, Fixed, 95% CI)	Subtotals only

1.1.1 Controlled‐release (CR) oxycodone vs immediate‐release (IR) oxycodone	3	319	Std. Mean Difference (IV, Fixed, 95% CI)	0.12 [‐0.10, 0.34]
1.1.2 CR oxycodone vs extended‐release (ER) oxycodone	1	62	Std. Mean Difference (IV, Fixed, 95% CI)	0.08 [‐0.42, 0.57]
1.1.3 CR oxycodone vs CR morphine	10	1137	Std. Mean Difference (IV, Fixed, 95% CI)	‐0.18 [‐0.30, ‐0.06]
1.1.4 CR oxycodone vs CR hydromorphone	1	62	Std. Mean Difference (IV, Fixed, 95% CI)	‐0.13 [‐0.63, 0.37]
1.1.5 CR oxycodone vs ER hydromorphone	2	259	Std. Mean Difference (IV, Fixed, 95% CI)	0.04 [‐0.21, 0.28]
1.1.6 CR oxycodone vs ER oxymorphone	1	74	Std. Mean Difference (IV, Fixed, 95% CI)	0.23 [‐0.23, 0.69]
1.1.7 CR oxycodone vs ER tapentadol	1	265	Std. Mean Difference (IV, Fixed, 95% CI)	‐0.06 [‐0.30, 0.18]
1.1.8 CR oxycodone vs transdermal (TD) fentanyl	2	329	Std. Mean Difference (IV, Fixed, 95% CI)	0.02 [‐0.19, 0.24]
1.1.9 CR oxycodone vs TD buprenorphine	1	252	Std. Mean Difference (IV, Fixed, 95% CI)	0.10 [‐0.15, 0.35]
1.1.10 IR oxycodone vs IR morphine	1	38	Std. Mean Difference (IV, Fixed, 95% CI)	‐0.15 [‐0.79, 0.49]
1.1.11 IR oxycodone vs IR hydromorphone	1	172	Std. Mean Difference (IV, Fixed, 95% CI)	0.15 [‐0.15, 0.45]
1.1.12 IV oxycodone vs IV morphine	1	65	Std. Mean Difference (IV, Fixed, 95% CI)	‐0.17 [‐0.66, 0.31]
1.2 Complete and/or significant pain relief (categorical) Show forest plot	17		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only

1.2.1 CR oxycodone vs CR morphine	13	1249	Risk Ratio (M‐H, Fixed, 95% CI)	1.02 [0.95, 1.10]
1.2.2 CR oxycodone vs ER tapentadol	1	265	Risk Ratio (M‐H, Fixed, 95% CI)	0.85 [0.65, 1.10]
1.2.3 CR oxycodone vs oral ibuprofen	1	66	Risk Ratio (M‐H, Fixed, 95% CI)	1.95 [1.24, 3.07]
1.2.4 CR oxycodone vs transdermal (TD) fentanyl	2	329	Risk Ratio (M‐H, Fixed, 95% CI)	0.98 [0.85, 1.14]
1.2.5 CR oxycodone vs TD buprenorphine	1	252	Risk Ratio (M‐H, Fixed, 95% CI)	0.94 [0.82, 1.09]
1.2.6 IV oxycodone vs IV morphine	1	65	Risk Ratio (M‐H, Fixed, 95% CI)	1.02 [0.70, 1.48]

Comparison 1. Pain

Comparison 2. Adverse events

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
2.1 Any adverse events Show forest plot	10		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only

2.1.1 Controlled‐release (CR) oxycodone vs CR morphine	8	943	Risk Ratio (M‐H, Fixed, 95% CI)	0.77 [0.70, 0.85]
2.1.2 CR oxycodone vs extended‐release (ER) hydromorphone	2	434	Risk Ratio (M‐H, Fixed, 95% CI)	1.06 [0.98, 1.14]
2.2 Severe adverse events Show forest plot	2		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only

2.2.1 CR oxycodone vs ER hydromorphone	2	434	Risk Ratio (M‐H, Fixed, 95% CI)	1.08 [0.73, 1.62]
2.3 Serious adverse events, including death Show forest plot	2		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only

2.3.1 CR oxycodone vs ER hydromorphone	2	434	Risk Ratio (M‐H, Fixed, 95% CI)	1.44 [0.86, 2.39]
2.4 Appetite loss/anorexia Show forest plot	5		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only

2.4.1 CR oxycodone vs CR morphine	3	263	Risk Ratio (M‐H, Fixed, 95% CI)	1.20 [0.36, 3.94]
2.4.2 CR oxycodone vs ER hydromorphone	2	434	Risk Ratio (M‐H, Fixed, 95% CI)	0.89 [0.54, 1.49]
2.5 Asthenia Show forest plot	2		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only

2.5.1 CR oxycodone vs immediate‐release (IR) oxycodone	2	208	Risk Ratio (M‐H, Fixed, 95% CI)	0.58 [0.20, 1.68]
2.6 Confusion Show forest plot	5		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only

2.6.1 CR oxycodone vs IR oxycodone	2	157	Risk Ratio (M‐H, Fixed, 95% CI)	0.78 [0.20, 3.02]
2.6.2 CR oxycodone vs CR morphine	3	584	Risk Ratio (M‐H, Fixed, 95% CI)	1.01 [0.78, 1.31]
2.7 Constipation Show forest plot	24		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only

2.7.1 CR oxycodone vs IR oxycodone	3	317	Risk Ratio (M‐H, Fixed, 95% CI)	0.71 [0.45, 1.13]
2.7.2 CR oxycodone vs CR morphine	18	1894	Risk Ratio (M‐H, Fixed, 95% CI)	0.75 [0.66, 0.86]
2.7.3 CR oxycodone vs transdermal (TD) fentanyl	2	336	Risk Ratio (M‐H, Fixed, 95% CI)	1.07 [0.88, 1.32]
2.7.4 CR oxycodone vs ER hydromorphone	2	434	Risk Ratio (M‐H, Fixed, 95% CI)	1.10 [0.80, 1.49]
2.8 Diarrhoea Show forest plot	2		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only

2.8.1 CR oxycodone vs ER hydromorphone	2	434	Risk Ratio (M‐H, Fixed, 95% CI)	0.91 [0.55, 1.49]
2.9 Dizziness/lightheadedness Show forest plot	16		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only

2.9.1 CR oxycodone vs IR oxycodone	3	317	Risk Ratio (M‐H, Fixed, 95% CI)	0.74 [0.40, 1.37]
2.9.2 CR oxycodone vs CR morphine	11	941	Risk Ratio (M‐H, Fixed, 95% CI)	0.87 [0.58, 1.31]
2.9.3 CR oxycodone vs ER hydromorphone	2	434	Risk Ratio (M‐H, Fixed, 95% CI)	1.00 [0.61, 1.64]
2.10 Drowsiness/somnolence Show forest plot	18		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only

2.10.1 CR oxycodone vs IR oxycodone	3	317	Risk Ratio (M‐H, Fixed, 95% CI)	1.03 [0.69, 1.54]
2.10.2 CR oxycodone vs CR morphine	15	1486	Risk Ratio (M‐H, Fixed, 95% CI)	0.88 [0.74, 1.05]
2.11 Dry mouth Show forest plot	8		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only

2.11.1 CR oxycodone vs IR oxycodone	3	317	Risk Ratio (M‐H, Fixed, 95% CI)	1.14 [0.48, 2.75]
2.11.2 CR oxycodone vs CR morphine	5	888	Risk Ratio (M‐H, Fixed, 95% CI)	0.97 [0.78, 1.22]
2.12 Dysuria/uroschesis Show forest plot	8		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only

2.12.1 CR oxycodone vs CR morphine	7	887	Risk Ratio (M‐H, Fixed, 95% CI)	0.64 [0.38, 1.07]
2.12.2 CR oxycodone vs TD fentanyl	2	336	Risk Ratio (M‐H, Fixed, 95% CI)	1.15 [0.62, 2.16]
2.13 Fever Show forest plot	2		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only

2.13.1 CR oxycodone vs ER hydromorphone	2	434	Risk Ratio (M‐H, Fixed, 95% CI)	1.04 [0.66, 1.62]
2.14 Hallucinations Show forest plot	4		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only

2.14.1 CR oxycodone vs CR morphine	4	696	Risk Ratio (M‐H, Fixed, 95% CI)	0.52 [0.28, 0.97]
2.15 Headache Show forest plot	3		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected

2.15.1 CR oxycodone vs IR oxycodone	3		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected
2.16 Insomnia Show forest plot	2		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only

2.16.1 CR oxycodone vs IR oxycodone	2	269	Risk Ratio (M‐H, Fixed, 95% CI)	1.04 [0.31, 3.53]
2.17 Insomnia & lethargy Show forest plot	2		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only

2.17.1 CR oxycodone vs CR morphine	2	314	Risk Ratio (M‐H, Fixed, 95% CI)	0.48 [0.26, 0.90]
2.18 Nausea Show forest plot	18		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only

2.18.1 CR oxycodone vs IR oxycodone	3	317	Risk Ratio (M‐H, Fixed, 95% CI)	0.85 [0.56, 1.28]
2.18.2 CR oxycodone vs CR morphine	13	1388	Risk Ratio (M‐H, Fixed, 95% CI)	0.93 [0.77, 1.12]
2.18.3 CR oxycodone vs ER hydromorphone	2	434	Risk Ratio (M‐H, Fixed, 95% CI)	0.83 [0.63, 1.08]
2.19 Nausea & vomiting Show forest plot	6		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only

2.19.1 CR oxycodone vs CR morphine	6	637	Risk Ratio (M‐H, Fixed, 95% CI)	0.77 [0.56, 1.06]
2.20 Nervousness Show forest plot	2		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only

2.20.1 CR oxycodone vs IR oxycodone	2	208	Risk Ratio (M‐H, Fixed, 95% CI)	0.57 [0.20, 1.64]
2.21 Pruritus Show forest plot	11		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only

2.21.1 CR oxycodone vs IR oxycodone	3	317	Risk Ratio (M‐H, Fixed, 95% CI)	1.46 [0.65, 3.25]
2.21.2 CR oxycodone vs CR morphine	8	1108	Risk Ratio (M‐H, Fixed, 95% CI)	0.76 [0.51, 1.14]
2.22 Sweating Show forest plot	4		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only

2.22.1 CR oxycodone vs IR oxycodone	2	269	Risk Ratio (M‐H, Fixed, 95% CI)	0.65 [0.22, 1.93]
2.22.2 CR oxycodone vs CR morphine	2	220	Risk Ratio (M‐H, Fixed, 95% CI)	4.52 [0.54, 37.94]
2.23 Vomiting Show forest plot	18		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only

2.23.1 CR oxycodone vs IR oxycodone	3	317	Risk Ratio (M‐H, Fixed, 95% CI)	0.66 [0.38, 1.15]
2.23.2 CR oxycodone vs CR morphine	13	1388	Risk Ratio (M‐H, Fixed, 95% CI)	0.81 [0.63, 1.04]
2.23.3 CR oxycodone vs ER hydromorphone	2	434	Risk Ratio (M‐H, Fixed, 95% CI)	0.84 [0.63, 1.10]
2.24 Discontinuation due to adverse events Show forest plot	12		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only

2.24.1 CR oxycodone vs IR oxycodone	3	317	Risk Ratio (M‐H, Fixed, 95% CI)	0.60 [0.29, 1.22]
2.24.2 CR oxycodone vs CR morphine	7	618	Risk Ratio (M‐H, Fixed, 95% CI)	0.79 [0.36, 1.73]
2.24.3 CR oxycodone vs ER hydromorphone	2	434	Risk Ratio (M‐H, Fixed, 95% CI)	1.09 [0.69, 1.75]

Comparison 2. Adverse events